SOLAR RESOURCE AND PV POTENTIAL OF THE MALDIVES 12 MONTH SOLAR RESOURCE REPORT June 2017 This report was prepared by Solargis, under contract to The World Bank. It is one of several outputs from the solar resource mapping component of the activity Energy Resource Mapping and Geospatial Planning Maldives [Project ID: P146018]. This activity is funded and supported by the Energy Sector Management Assistance Program (ESMAP), a multi-donor trust fund administered by The World Bank, under a global initiative on Renewable Energy Resource Mapping. Further details on the initiative can be obtained from the ESMAP website. This document is an interim output from the above-mentioned project, and the content is the sole responsibility of the consultant authors. Users are strongly advised to exercise caution when utilizing the information and data contained, as this may include preliminary data and/or findings, and the document has not been subject to full peer review. Final outputs from this project will be marked as such, and any improved or validated solar resource data will be incorporated into the Global Solar Atlas. Copyright © 2018 THE WORLD BANK Washington DC 20433 Telephone: +1-202-473-1000 Internet: www.worldbank.org The World Bank, comprising the International Bank for Reconstruction and Development (IBRD) and the International Development Association (IDA), is the commissioning agent and copyright holder for this publication. However, this work is a product of the consultants listed, and not of World Bank staff. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work and accept no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for non-commercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: +1-202-522-2625; e-mail: pubrights@worldbank.org. Furthermore, the ESMAP Program Manager would appreciate receiving a copy of the publication that uses this publication for its source sent in care of the address above, or to esmap@worldbank.org. Annual Solar Resource Report for solar meteorological stations after completion of 12 months of measurements Republic of Maldives Reference No. 129-06/2017 Date: 7 June 2017 Customer Consultant World Bank Solargis s.r.o. Energy Sector Management Assistance Program Contact: Mr. Marcel Suri Contact: Mr. Oliver J. Knight Mytna 48, 811 07 Bratislava, Slovakia 1818 H St NW, Washington DC, 20433, USA Phone +421 2 4319 1708 Phone: +1-202-473-3159 E-mail: marcel.suri@solargis.com E-mail: mailto:oknight@worldbank.org http://solargis.com http://www.esmap.org/RE_Mapping Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 TABLE OF CONTENTS Table of contents ...................................................................................................................................................... 4 Acronyms .................................................................................................................................................................. 6 Glossary .................................................................................................................................................................... 7 1 Introduction ...................................................................................................................................................... 8 1.1 Background ........................................................................................................................................... 8 1.1 Delivered data sets................................................................................................................................ 8 1.2 Information included in this report ........................................................................................................ 9 2 Position of solar meteorological sites ........................................................................................................... 10 3 Ground measurements in Maldives ............................................................................................................... 12 3.1 Instruments and measured parameters .............................................................................................. 12 3.2 Station operation and calibration of instruments ................................................................................ 13 3.3 Quality control of measured solar resource data ................................................................................ 14 3.3.1 Hanimaadhoo Airport ............................................................................................................ 14 3.3.2 Hulhulé Airport....................................................................................................................... 16 3.3.3 Kadhdhoo Airport .................................................................................................................. 18 3.3.4 Gan Airport ............................................................................................................................ 20 3.4 Recommendations on the operation and maintenance of the sites .................................................... 23 4 Solar resource model data ............................................................................................................................. 24 4.1 Solar model ......................................................................................................................................... 24 4.2 Site adaptation of the solar model − method ...................................................................................... 25 4.3 Results of the model adaptation at four sites ..................................................................................... 26 5 Meteorological model data ............................................................................................................................ 33 5.1 Meteorological model ......................................................................................................................... 33 5.2 Validation of meteorological data ....................................................................................................... 33 5.2.1 Air temperature at 2 metres................................................................................................... 34 5.2.2 Relative humidity ................................................................................................................... 36 5.2.3 Wind speed and wind direction at 10 metres......................................................................... 39 5.3 Uncertainty of meteorological model data ........................................................................................... 41 6 Solar resource: uncertainty of longterm estimates ....................................................................................... 42 6.1 Uncertainty of solar resource yearly estimate ..................................................................................... 42 6.2 Uncertainty due to interannual variability of solar radiation ................................................................ 43 6.3 Combined uncertainty ......................................................................................................................... 44 7 Time series and Typical Meteorological Year data........................................................................................ 48 7.1 Delivered data sets.............................................................................................................................. 48 7.2 TMY method ....................................................................................................................................... 48 7.3 Results ................................................................................................................................................ 49 8 Conclusions .................................................................................................................................................... 53 Annex 1: Site related data statistics ....................................................................................................................... 54 Yearly summaries of solar and meteorological parameters ........................................................................... 54 Monthly summaries of solar and meteorological parameters ........................................................................ 56 © 2017 Solargis page 4 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Frequency of occurrence of GHI and DNI daily model values for a period 1999 to 2016................................ 58 Frequency of occurrence of GHI and DNI 30-minute model values for a period 1999 to 2016 ....................... 62 Frequency of occurrence of GHI and DNI measured and model values representing year 2016 .................... 66 List of figures .......................................................................................................................................................... 72 List of tables ........................................................................................................................................................... 74 References .............................................................................................................................................................. 76 Support information ................................................................................................................................................ 78 Background on Solargis .................................................................................................................................. 78 Legal information ........................................................................................................................................... 78 © 2017 Solargis page 5 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 ACRONYMS AP Atmospheric Pressure CFSR Climate Forecast System Reanalysis. The meteorological model operated by the US service NOAA (National Oceanic and Atmospheric Administration) CFS v2 Climate Forecast System Version 2 CFSv2 model is the operational extension of the CFSR (NOAA, NCEP) DIF Diffuse Horizontal Irradiation, if integrated solar energy is assumed. Diffuse Horizontal Irradiance, if solar power values are discussed DNI Direct Normal Irradiation, if integrated solar energy is assumed. Direct Normal Irradiance, if solar power values are discussed. GFS Global Forecast System. The meteorological model operated by the US service NOAA (National Oceanic and Atmospheric Administration) GHI Global Horizontal Irradiation, if integrated solar energy is assumed. Global Horizontal Irradiance, if solar power values are discussed. MACC Monitoring Atmospheric Composition and Climate – meteorological model operated by the European service ECMWF (European Centre for Medium-Range Weather Forecasts) Meteosat IODC (MFG) Meteosat satellites operated by EUMETSAT organization. MFG: Meteosat First Generation. For this report the data from the Meteosat IODC are used. The satellite is positioned over the Indian Ocean. PWAT Precipitable water (water vaour) RH Relative Humidity at 2 metres TEMP Air Temperature at 2 metres WD Wind Direction at 10 metres WS Wind Speed at 10 metres © 2017 Solargis page 6 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 GLOSSARY Aerosols Small solid or liquid particles suspended in air, for example desert sand or soil particles, sea salts, burning biomass, pollen, industrial and traffic pollution. All-sky irradiance The amount of solar radiation reaching the Earth's surface is mainly determined by Earth-Sun geometry (the position of a point on the Earth's surface relative to the Sun which is determined by latitude, the time of year and the time of day) and the atmospheric conditions (the level of cloud cover and the optical transparency of atmosphere). All-sky irradiance is computed with all factors taken into account Bias Represents systematic deviation (over- or underestimation) and it is determined by systematic or seasonal issues in cloud identification algorithms, coarse resolution and regional imperfections of atmospheric data (aerosols, water vapour), terrain, sun position, satellite viewing angle, microclimate effects, high mountains, etc. Clear-sky irradiance The clear sky irradiance is calculated similarly to all-sky irradiance but without taking into account the impact of cloud cover. Long-term average Average value of selected parameter (GHI, DNI, etc.) based on multiyear historical time series. Long-term averages provide a basic overview of solar resource availability and its seasonal variability. P50 value Best estimate or median value represents 50% probability of exceedance. For annual and monthly solar irradiation summaries it is close to average, since multiyear distribution of solar radiation resembles normal distribution. P90 value Conservative estimate, assuming 90% probability of exceedance (with the 90% probability the value should be exceeded). When assuming normal distribution, the P90 value is also a lower boundary of the 80% probability of occurrence. P90 value can be calculated by subtracting uncertainty from the P50 value. In this report, we apply a simplified assumption of normal distribution of yearly values. Root Mean Square Represents spread of deviations given by random discrepancies between measured Deviation (RMSD) and modelled data and is calculated according to this formula: 3 ' 2 '45 ' ()*+,-). − (0.)1). = On the modelling side, this could be low accuracy of cloud estimate (e.g. intermediate clouds), under/over estimation of atmospheric input data, terrain, microclimate and other effects, which are not captured by the model. Part of this discrepancy is natural - as satellite monitors large area (of approx. 3.3 x 3.4 km for MSG satellite), while sensor sees only micro area of approx. 1 sq. centimetre. On the measurement side, the discrepancy may be determined by accuracy/quality and errors of the instrument, pollution of the detector, misalignment, data loggers, insufficient quality control, etc. 2 Solar irradiance Solar power (instantaneous energy) falling on a unit area per unit time [W/m ]. Solar resource or solar radiation is used when considering both irradiance and irradiation. 2 Solar irradiation Amount of solar energy falling on a unit area over a stated time interval [Wh/m or 2 kWh/m ]. Uncertainty Is a parameter characterizing the possible dispersion of the values attributed to an of estimate, Uest estimated irradiance/irradiation values. In this report, uncertainty assessment of the solar resource model estimate is based on a detailed understanding of the achievable accuracy of the solar radiation model and its data inputs (satellite, atmospheric and other data), which is confronted by an extensive data validation experience. The second source of uncertainty is ground measurements. Their quality depends on accuracy of instruments, their maintenance and data quality control. Third contribution to the uncertainty is from the site adaptation method where ground-measured and satellite- based data are correlated. © 2017 Solargis page 7 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 1 INTRODUCTION 1.1 Background This report is prepared within Phase 1 of the project Renewable Energy Resource Mapping for the Republic of the Maldives. This part of the project focuses on solar resource mapping and measurement services as part of a technical assistance in the renewable energy development implemented by the World Bank in Maldives. It is being undertaken in close coordination with the Ministry of Environment and Energy (MEE) of Maldives, the World Bank’s primary country counterpart for this project. This project is funded by the Energy Sector Management Assistance Program (ESMAP) and Asia Sustainable and Alternative Energy Program (ASTAE), both administered by the World Bank and supported by bilateral donors. This report summarizes results of first 12 months of the measuring campaign by the project ESMAP at four solar meteorological stations, installed in Maldives as part of the World Bank’s ESMAP mission in Maldives. This report describes delivery of site-specific data, site adaptation of solar model, data uncertainty and statistical summary. This report accompanies delivery of site-specific solar resource and meteorological data for four sites, where solar meteorological stations have been operated. As a result of high-quality operation of the meteorological sits, and site adaptation the Solargis model, reliable historical time series and TMY data is computed. The delivered time series and Typical Meteorological Year data is ready for use for bankable evaluation of solar energy projects. The measurements at four sites, located in Maldives, are provided by Suntrace company (Germany). The model data for the same sites and related works, together with this report are supplied by Solargis (Slovakia). 1.1 Delivered data sets The site specific data, provided as part of this delivery, include: • Solar and meteorological easurements, after second level quality assessment (first level was delivered by Suntrace) • Time series, representing last 18 years (1999 to 2016) • Typical Meteorological Year data, also representing last 18 years The data is delivered in formats ready to use in energy simulation software. This report provides detailed insight of the methodologies and reuslts. Table 1.1 Delivered data characterstics Feature Time coverage Primary time step Delivered files Measurements Dec 2015 to Dec 2016 1 minute Quality controlled measurements – 1- minute (Suntrace) Analytical tasks consider only a period from Jan to Dec 2016 Model data – original Jan 1999 to Dec 2016 30 minutes Time series – hourly (Solargis) Time series – monthly Time series – yearly Model data – site dadapted Jan 1999 to Dec 2016 30 minutes Time series – hourly (Solargis) Time series – monthly Time series – yearly Typical Meteorological Year P50 – hourly Typical Meteorological Year P90 – hourly © 2017 Solargis page 8 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Table 1.2 Parameters in the delivered site-adapted time series and TMY data (hourly time step) Parameter Acronym Unit TS TMY P50 TMY P90 2 Global horizontal irradiance GHI W/m X X X 2 Direct normal irradiance DNI W/m X X X 2 Diffuse horizontal irradiance DIF W/m X X X 2 Global tilted irradiance (at optimum angle) GHI W/m X - - Solar azimuth SA ° X X X Solar elevation SE ° X X X Air temperature at 2 metres TEMP °C X X X Wind speed at 10 metres WS m/s X X X Wind direction at 10 metres WD ° X X X Relative humidity RH % X X X Air Pressure AP hPa X X X 2 Precipitable Water PWAT kg/m X X X 1.2 Information included in this report This report presents: • Solar resource and meteorological measurements after 12 months of operation o Review and quality check of the measured data o Calibration procedures and results o List and explanation of the occurred disturbances and failures • Comparison of the measurements to the model and uncertainty analysis o Comparison of solar and meteo measurements with the model data o Site adaptation of satellite data based on ground measurements and uncertainty estimate o Estimate of data uncertainty • Data analysis (measured vs. modelled) o Monthly summaries of solar and meteorological parameters captured at the site o Variability of measured solar parameters o Frequency of occurrence of GHI and DNI 1-minute and 30-minute values o Frequency of occurrence of GHI and DNI 1-minute and 30-minute ramps © 2017 Solargis page 9 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 2 POSITION OF SOLAR METEOROLOGICAL SITES In Maldives, four measuring stations were installed within the ESMAP Solar initiative. All of them have been located at the airports within the premises of Maldives Meteorological Service (Figure 2.1, Table 2.1). Figure 2.1: Position of solar meteorological stations in Maldives Table 2.1 Overview information on the solar meteorological stations installed in Maldives Site name Latitude Longitude Altitude (airport) Site ID [º] [º] [m a.s.l.] Measurement station host Hanimaadhoo MVHAQ 6.7482° 73.1696° 2 Hanimaadhoo International Airport Hulhulé MVMLE 4.1927° 73.5281° 2 Male International Airport Kadhdhoo MVKDO 1.8599° 73.5203° 2 Kadhdhoo Airport Gan MVGAN -0.6911° 73.1599° 2 Gan International Airport © 2017 Solargis page 10 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Besides the good geographical distribution within the territory of Maldives, the locations fit well to the population centres, where solar installations can be primarily deployed. In addition to geographical and socio-economic criteria, the sites fulfil the criteria for the operation and maintenance of the solar measuring stations: • Availability of free horizon, • Availability of GSM networks, • Availability of local work force for maintenance, • Easy to access and high level of security During the first year of the measurements, the measured data was analysed and harmonized with the objective to acquire reference solar radiation data for reducing the uncertainty of the model Chapter 6). The quality data from four Tier-2 meteorological stations are available for this assessment (Chapter 3). Position and detailed information about measurement sites is available also on Global Solar Atlas website: http://globalsolaratlas.info/?c=4.850154,73.146973,6&e=1 © 2017 Solargis page 11 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 3 GROUND MEASUREMENTS IN MALDIVES 3.1 Instruments and measured parameters Basic information about measurements sites is in Table 3.1. Solar parameters at all stations are measured by high (SR20, for measurements of GHI) and medium (RSP, for GHI, DNI and DIF) accuracy equipment (Table 3.2 and 3.3). The measurement campaign in Maldives has been performed by Suntrace GmbH company (Germany). Table 3.1 Overview information on measurement stations operated in the region Site name Latitude Longitude Altitude ID (airport) Site ID [º] [º] [m a.s.l.] Installation date 1 Hanimaadhoo MVHAQ 6.7482° 73.1696° 2 10 Dec 2015 2 Hulhulé MVMLE 4.1927° 73.5281° 2 08 Dec 2015 3 Kadhdhoo MVKDO 1.8599° 73.5203° 2 14 Dec 2015 4 Gan MVGAN -0.6911° 73.1599° 2 13 Dec 2015 Table 3.2 Instruments installed at the solar meteorological stations Site name GHI 1 GHI 2 DNI DIF Temp RH WS AP Hanimaadhoo SR20 - T1 RSP 4G RSP 4G RSP 4G DKRF 400 DKRF 400 First Class - Hulhulé SR20 - T1 RSP 4G RSP 4G RSP 4G DKRF 400 DKRF 400 First Class - Kadhdhoo SR20 - T1 RSP 4G RSP 4G RSP 4G DKRF 400 DKRF 400 First Class - Gan SR20 - T1 RSP 4G RSP 4G RSP 4G DKRF 400 DKRF 400 First Class - Table 3.3 Technical parameters and accuracy class of the instruments Parameter Instrument Type Manufacturer Uncertainty GHI 1 Secondary standard pyranometer SR20 - T1 Hukseflux < ±2.0 % (daily) GHI 2 Rotating Shadowband Irradiometer RSP 4G Reichert GmbH Indicatively ±5 % DIF Rotating Shadowband Irradiometer RSP 4G Reichert GmbH Indicatively ±8 % DNI Rotating Shadowband Irradiometer RSP 4G Reichert GmbH Indicatively ±5 % TEMP Temperature probe DKRF 400 Driesen und Kern ± 1.5 ºC RH Relative humidity probe DKRF 400 Driesen und Kern ± 3.5 % RH WS Wind speed sensor (WS at 3.0 m height) First Class Thies Clima < ±1 % AP Barometric pressure sensor Not provided - - - Data logger blueberry Wilmers GmbH - © 2017 Solargis page 12 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 3.2 Station operation and calibration of instruments In this report, data from the first year of 2-year measurement campaign is analysed. As the measurement stations have been installed during December 2015, the period considered for the data analysis starts in January 2016 and ends in December 2016 for all four stations. Overview of the data availability, time step and measured parameters is shown in Tables 3.4 and 3.5. Table 3.4 Overview information on solar meteorological stations operating in the region Site name(airport) Site ID First measurement period Primary time step Hanimaadhoo MVHAQ 11 Dec 2015 to 31 Dec 2016 1 minute Hulhulé MVMLE 09 Dec 2015 to 31 Dec 2016 1 minute Kadhdhoo MVKDO 15 Dec 2015 to 31 Dec 2016 1 minute Gan MVGAN 14 Dec 2015 to 31 Dec 2016 1 minute Table 3.5 Period of measurements analysed in this report Year, month 2015 2016 2017 Station 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Hanimaadhoo Hulhulé Kadhdhoo Gan 1 min time step of measurements During the measurement campaign, local staff of Maldives Meteorological Service was fully trained by Suntrace, and participating on daily inspection and cleaning of instruments. Local partner Renewable Energy Maldives, performed detailed visits and station maintenance after 6 and 12 months of operation. Instruments field verification, i.e. one day long comparative measurements of solar radiation parameters and cross check with the instruments from the spare station to proof that sensitivities (calibration constants) remained stable within the instrument specifications, was performed by Suntrace GmbH after one year of operation (Table 3.6). Table 3.6 Meteorological stations maintenance and instruments field verification Site name Instruments Instruments field Site ID Annual verification visit Comments (airport) cleaning interval verification SR20 - T1 Hanimaadhoo MVHAQ daily 23 to 24 Nov 2016 No issues RSP 4G SR20 - T1 Hulhulé MVMLE daily 22 to 23 Nov 2016 No issues RSP 4G SR20 - T1 Kadhdhoo MVKDO daily 27 to 28 Nov 2016 No issues RSP 4G SR20 - T1 Gan MVGAN daily 25 to 26 Nov 2016 No issues RSP 4G To perform a verification of the measurements, the spare instruments were connected to the spare station logger. The timestamp of both data has been synchronized to be able to perform a valid inter-comparison. For 2 the RSP values the corresponding calibration factors and corrections have been applied. Values below 50 W/m have been discarded. Results of field instruments verification are listed in Table 3.7. According to the results of filed check of the instruments, the calibration coefficients at Gan, Kadhdhoo and Hulhulé solar meteorological © 2017 Solargis page 13 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 stations has been updated. Detailed results and discussion is supplied in Annual maintenance visit reports delivered separately for each meteorological station in February 2017. Table 3.7 Results of field instruments verification at the respective stations Relative bias [%] Site name Site ID SR20 RSP 4G RSP 4G RSP 4G (airport) GHI GHI DNI DIF Hanimaadhoo MVHAQ -0.7 2.6 14.8* 1.2 Hulhulé MVMLE -2.3 -3.5 4.5 -5.5 Kadhdhoo MVKDO -0.8 0.8 5.7 -2.0 Gan MVGAN 0.0 2.3 15.2* -0.6 * Verification affected by prevailing cloud weather 3.3 Quality control of measured solar resource data Prior to correlation with satellite-based solar data, the ground-measured solar radiation was quality-controlled by Solargis. Quality control (QC) was based on methods defined in SERI QC procedures and Younes et al. [1, 2] and the in-house developed tests. The ground measurements were inspected also visually, mainly for identification of shading and other error patterns such as RSP shading ring malfunction. Figures 3.1 to 3.7 show results of quality control for individual stations. The colours in Figures 3.1, 3.3, 3.5 and 3.7 indicate the following flags: • Blue: data excluded by visual inspection - mainly shading, shading ring issues • Green: data passed all tests • Grey: sun below horizon • White and brown strips: missing data • Red and violet: GHI, DNI and DIF consistency problem or problems with physical limits The data records not passing the quality control test were flagged and excluded from the further processing. The results show relatively small amount of excluded data readings (Tables 3.9, 3.11, 3.13 and 3.15). Most serious is the DNI degradation due to shadowband issues. 3.3.1 Hanimaadhoo Airport Table 3.8 Occurrence of data readings for Hanimaadhoo meteorological station Data availability GHI SR20 GHI, DNI RSP Sun below horizon 276 509 49.6% 276 509 49.6% Sun above horizon 280 770 50.4% 280 770 50.4% Total data readings 557 279 100.0% 557 279 100.0% © 2017 Solargis page 14 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Table 3.9 Excluded ground measurements after quality control (Sun above horizon) in Hanimaadhoo Occurrence of data samples (Sun above horizon) Type of test GHI SR20 DNI RSP GHI RSP Physical limits test 237 0.1% 0 0.0% 1 288 0.5% Consistency test (GHI – DNI – DIF) - - 135 0.0% 135 0.0% Visual test (incorrect data) 12 886 4.6% 17 995 6.4% 12 462 4.4% Other (non valid data) 98 0.0% 166 0.1% 166 0.1% Total excluded data samples 13 221 4.7% 18 296 6.5% 14 051 5.0% Total samples 280 770 100.0% 280 770 100.0% 280 770 100.0% Figure 3.1 Results of GHI and DNI quality control in Hanimaadhoo. Green – data passing all tests; grey – sun below horizon; red – consistency issue, violet – physical limit, blue excluded by visual inspection. Top: DNI (RSP); middle: GHI (RSP); bottom: GHI (SR20) © 2017 Solargis page 15 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Main findings: • Shadowband malfunction. Due to specific location of measurement stations close to the Equator the rotating shadowband may not in some periods sufficiently shade the sensor. This insufficient shading results in degraded DNI and DIF measurements. This problem was identified for period a from 3 to 26 June 2016 around the noon. Approximately 1.8% of DNI and DIF measurements were excluded. • A systematic difference between GHI measurements from secondary standard pyranometer SR20-T1 and RSP (Figure 3.4). GHI from SR20 is in average higher by 4.4% than GHI from RSP. In the noon time, it can exceed 6%. • Late afternoon shading from surrounding objects • GHI measurements from SR20 were replaced by data from RSP in case of shading from wind mast. Figure 3.2 Systematic difference between GHI from SR20 and RSP − Hanimaadhoo. 3.3.2 Hulhulé Airport Table 3.10 Occurrence of data readings for Hulhulé meteorological station Data availability GHI SR20 GHI, DNI RSP Sun below horizon 277 856 49.6% 277 856 49.6% Sun above horizon 282 303 50.4% 282 303 50.4% Total data readings 560 159 100.0% 560 159 100.0% Table 3.11 Excluded ground measurements after quality control (Sun above horizon) in Hulhulé Occurrence of data samples (Sun above horizon) Type of test GHI SR20 DNI RSP GHI RSP Physical limits test 278 0.1% 0 0.0% 1 211 0.4% Consistency test (GHI – DNI – DIF) - - 52 0.0% 52 0.0% Visual test (incorrect data) 12 796 4.5% 24 692 8.7% 12 436 4.4% Other (non valid data) 101 0.0% 5 687 2.0% 853 0.3% Total excluded data samples 13 175 4.7% 30 431 10.8% 14 552 5.2% Total samples 282 303 100.0% 282 303 100.0% 282 303 100.0% © 2017 Solargis page 16 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.3 Results of GHI and DNI quality control in Hulhulé. Green – data passing all tests; grey – sun below horizon; red – consistency issue, violet – physical limit issue, blue excluded by visual inspection; brown – missing data. Top: DNI (RSP); middle: GHI (RSP); bottom: GHI (SR20) Main findings: • Shadowband malfunction (Figure 3.4). This issue was identified for a period from 15 May to 19 July 2016 around the noon. Affected data readings in May were excluded by data provider, readings from June and July were excluded by quality control. Approximately 6.2% of DNI and DIF measurements were excluded in total. • A systematic difference between GHI measurements from secondary standard pyranometer SR20-T1 and RSP (Figure 3.5). GHI from SR20 is in average higher by 6.6% than GHI from RSP. In the noontime the difference can exceed 8%. • Late afternoon shading from surrounding objects • GHI measurements from SR20 were replaced by data from RSP in case of shading from wind mast. © 2017 Solargis page 17 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.4 Effect of RSP shadowband issues – drop of DNI in Hulhulé. Blue: DNI RSP; yellow: GHI SR20; red: GHI RSP; green: DIF RSP; dashed: theoretical clear-sky profile Figure 3.5 Systematic difference between GHI from SR20 and RSP - Hulhulé. 3.3.3 Kadhdhoo Airport Table 3.12 Occurrence of data readings for Kadhdhoo meteorological station Data availability GHI SR20 GHI, DNI RSP Sun below horizon 273 387 49.6% 273 387 49.6% Sun above horizon 278 132 50.4% 278 132 50.4% Total data readings 551 519 100.0% 551 519 100.0% Table 3.13 Excluded ground measurements after quality control (Sun above horizon) in Kadhdhoo Occurrence of data samples (Sun above horizon) Type of test GHI SR20 DNI RSP GHI RSP Physical limits test 381 0.1% 0 0.0% 1 089 0.4% Consistency test (GHI – DNI – DIF) - - 66 0.0% 66 0.0% Visual test (incorrect data) 23 762 8.5% 41 994 15.1% 23 011 8.3% Other (non valid data) 48 0.0% 102 0.0% 101 0.0% Total excluded data samples 24 191 8.7% 42 162 15.2% 24 267 8.7% Total samples 278 132 100.0% 278 132 100.0% 278 132 100.0% © 2017 Solargis page 18 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.6 Results of GHI and DNI quality control − Kadhdhoo. Green – data passing all tests; grey – sun below horizon; red – consistency issue, violet – physical limit issue, blue excluded by visual inspection. Top: DNI (RSP); middle: GHI (RSP); bottom: GHI (SR20) Main findings: • Shadowband malfunction (Figure 3.7). This issue was identified for a period from 25 May to 7 July 2016 around the noon. Affected data readings from June and July were excluded by quality control. Approximately 6.8% of DNI and DIF measurements were excluded in total. • A systematic difference between GHI measurements from secondary standard pyranometer SR20-T1 and RSP (Figure 3.8). GHI from SR20 is in average higher by 4.3% than GHI from RSP. In the noon time it can exceed 7%. • Late afternoon and early morning shading from surrounding objects • GHI measurements from SR20 were replaced by data from RSP in case of shading from wind mast. © 2017 Solargis page 19 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.7 Effect of RSP shadowband issues – drop of DNI in Kadhdhoo. Blue: DNI RSP; yellow: GHI SR20; red: GHI RSP; green: DIF RSP; dashed: theoretical clear-sky profile Figure 3.8 Systematic difference between GHI from SR20 and RSP − Kadhdhoo. 3.3.4 Gan Airport Table 3.14 Occurrence of data readings for Gan meteorological station Data availability GHI SR20 GHI, DNI RSP Sun below horizon 273 987 49.5% 273 987 49.5% Sun above horizon 278 972 50.5% 278 972 50.5% Total data readings 552 959 100.0% 552 959 100.0% Table 3.15 Excluded ground measurements after quality control (Sun above horizon) in Gan Occurrence of data samples (Sun above horizon) Type of test GHI SR20 DNI RSP GHI RSP Physical limits test 393 0.1% 0 0.0% 1 168 0.4% Consistency test (GHI – DNI – DIF) - - 144 0.1% 144 0.1% Visual test (incorrect data) 0 0.0% 14 843 5.3% 0 0.0% Other (non valid data) 98 0.0% 17 194 6.2% 328 0.1% Total excluded data samples 491 0.2% 32 181 11.5% 1 640 0.6% Total samples 278 972 100.0% 278 972 100.0% 278 972 100.0% © 2017 Solargis page 20 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.9 Results of GHI and DNI quality control − Gan. Green – data passing all tests; grey – sun below horizon; red – consistency issue, violet – physical limit issue, blue excluded by visual inspection; brown – missing data. Top: DNI (RSP); middle: GHI (RSP); bottom: GHI (SR20) Main findings: • Shadowband malfunction (Figure 3.10). This issue was identified for several periods (from 14 December 2015 to 12 January 2016; from 1 May to 31 May 2016 and from 22 October to 25 November 2016) around the noon. Affected data readings were excluded partly by data provider and partly by quality control. Approximately 11% of DNI and DIF measurements were excluded in total. • A systematic difference between GHI measurements from secondary standard pyranometer SR20-T1 and RSP (Figure 3.11). GHI from SR20 is in average higher by 4.5% than GHI from RSP. In the noon time it can exceed 6%. • GHI measurements from SR20 were replaced by data from RSP in case of shading from wind mast (Figure 3.12). © 2017 Solargis page 21 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 3.10 Effect of RSP shadowband issues – drop of DNI in Gan. Blue: DNI RSP; yellow: GHI SR20; red: GHI RSP; green: DIF RSP; dashed: theoretical clear-sky profile Figure 3.11 Systematic difference between GHI from SR20 and RSP - Gan. Figure 3.12 Replacement of SR20 GHI by RSP GHI due to wind mast shading in Gan. Blue: DNI RSP; yellow: GHI SR20; red: GHI RSP; green: DIF RSP; dashed: theoretical clear-sky profile © 2017 Solargis page 22 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 3.4 Recommendations on the operation and maintenance of the sites Based on the results of quality control (Table 3.16), we conclude that the solar radiation measurements come from the high (SR20) and medium (RSP) accuracy equipment that is professionally operated and maintained. Some issues were identified during the data quality control: • Degraded DNI and DIF measurements due to shadowband malfunction. These data were flagged and excluded from further processing • Higher systematic difference between GHI from SR20 and RSP (up to 6.6% in Hulhulé). This is known effect that should relate only to GHI instrument calibration. Due to this issue, the GHI from RSP was not considered for site-adaptation For future works, we recommend: • Adopt measures to avoid the shadowband problems, where possible. • Identify and fix source of systematic difference between SR20 and RSP Table 3.16 Quality control summary (all sites) Description Station description, metadata Installation report for all stations available Instrument accuracy Secondary standard instrument SR20-T1 Rotating Shadowband Irradiometer RSP 4G Instrument calibration Instruments were calibrated Data structure Clear Cleaning and maintenance Diligent daily cleaning information Time reference Correct and clear time reference Quality control complexity RSP data, full QC SR20 data, without (GHI-DNI-DIF) consistency test Quality control results Shadowband temporary technical issues Larger systematic difference between GHI from SR20 and RSP (this is known issue) Period Almost 13 months Other issues Legend: Quality marker Very good Good Sufficient Problematic Insufficient Not specified © 2017 Solargis page 23 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 4 SOLAR RESOURCE MODEL DATA 4.1 Solar model Solar radiation is calculated by Solargis model, which are parameterized by a set of inputs characterizing the cloud transmittance, state of the atmosphere and terrain conditions. A comprehensive overview of the Solargis model is made available in the recent book publication [3]. The methodology is also described in [4, 5]. The related uncertainty and requirements for bankability are discussed in [6, 7]. In Solargis approach, the clear-sky irradiance is calculated by the simplified SOLIS model [8]. This model allows fast calculation of clear-sky irradiance from the set of input parameters. Sun position is deterministic parameter, and it is described by the algorithms with satisfactory accuracy. Stochastic variability of clear-sky atmospheric conditions is determined by changing concentrations of atmospheric constituents, namely aerosols, water vapour and ozone. Global atmospheric data, representing these constituents, are routinely calculated by world atmospheric data centres: • In Solargis, the new generation aerosol data set representing Atmospheric Optical Depth (AOD) is used. The calculation accuracy is strongly determined by quality of aerosols, especially for cloudless conditions. The aerosol data implemented by MACC-II/CAMS and MERRA-2 projects are used [9, 10]. • Water vapour is also highly variable in space and time, but it has lower impact on the values of solar radiation, compared to aerosols. The GFS and CFSR databases (NOAA NCEP) are used in Solargis, and the data represent the daily variability from 1999 to the present time [11, 12, 13]. • Ozone absorbs solar radiation at wavelengths shorter than 0.3 µm, thus having negligible influence on the broadband solar radiation. The clouds are the most influencing factor, modulating clear-sky irradiance. Effect of clouds is calculated from the satellite data in the form of a cloud index (cloud transmittance). The cloud index is derived by relating radiance recorded by the satellite in spectral channels and surface albedo to the cloud optical properties. In Solargis, the modified calculation scheme of Cano has been adopted to retrieve cloud optical properties from the satellite data [14]. To calculate all-sky irradiance in each time step, the clear-sky global horizontal irradiance is coupled with cloud index. Direct Normal Irradiance (DNI) is calculated from Global Horizontal Irradiance (GHI) using modified Dirindex model [15]. Diffuse irradiance for tilted surfaces, which is calculated by Perez model [16]. The calculation procedure includes also terrain disaggregation, the spatial resolution is enhanced with use of the digital terrain model to 250 meters [17]. Solargis model version 2.1 has been used. Table 4.1 summarizes technical parameters of the model inputs and of the primary data outputs. Table 4.1 Input data used in the Solargis and related GHI and DNI outputs for Maldives Inputs into the Solargis model Source Time Original Approx. grid of input data representation time step resolution Cloud index Meteosat IODC satellites 1999 to date 30 minutes 2.8 x 3.2 km (EUMETSAT) Atmospheric optical depth MACC-II/CAMS* 2003 to date 3 hours 75 km and 125 km (aerosols)* (ECMWF) MERRA-2 (NASA) 1999 to 2002 1 hour 50 km Water vapour CFSR/GFS 1999 to date 1 hour 35 and 55 km (NOAA) Elevation and horizon SRTM-3 - - 250 m (SRTM) Solargis primary data outputs - 1999 to date 30 minutes 250 m (GHI and DNI) * Aerosol data for 2003-2012 come from the reanalysis database; the data representing years 2013-present are derived from near-real time operational model © 2017 Solargis page 24 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 4.2 Site adaptation of the solar model − method The fundamental difference between a satellite observation and a ground measurement is that signal received by the satellite radiometer integrates a large area, while a ground station represents a pinpoint measurement. This results in a mismatch when comparing instantaneous values from these two observation instruments, mainly during intermittent cloudy weather and changing aerosol load. Nearly half of the hourly Root Mean Square Deviation (RMSD) for GHI and DNI can be attributed to this mismatch (value at sub-pixel scale), which is also known as the “nugget effect” [18]. The satellite pixel is not capable describing the inter-pixel variability in complex regions, where within one pixel diverse natural conditions mix-up (e.g. fog in narrow valleys or along the coast). In addition, the coarse spatial resolution of atmospheric databases such as aerosols or water vapour is not capable to describe local patterns of the state of atmosphere. These features can be seen in the satellite DNI data by increased bias due to imperfect description of aerosol load. Satellite data have inherent inaccuracies, which have certain degree of geographical and time variability. Especially DNI is strongly sensitive to variability of cloud information, aerosols, water vapour, and terrain shading. The relation between uncertainty of global and direct irradiance is nonlinear. Often, a negligible error in global irradiance may have high counterpart in the direct irradiance component. The solar energy projects require representative and accurate GHI and DNI time series. The satellite-derived databases are used to describe long-term solar resource for a specific site. However, their problem when compared to the high-quality ground measurements is a slightly higher bias and partial disagreement of frequency distribution functions, which may limit their potential to record the occurrence of extreme situations (e.g. very low atmospheric turbidity resulting in a high DNI and GHI). A solution is to correlate satellite-derived data with ground measurements to understand the source of discrepancy and subsequently to improve the accuracy of the resulting time series. The Solargis satellite-derived data are correlated with ground measurement data with two objectives: • Improvement of the overall bias (removal of systematic deviations) • Improvement of the fit of the frequency distribution of values. The relation between uncertainty of global and direct irradiance is nonlinear. Often, a negligible error in the global horizontal irradiance may trigger higher error in the direct irradiance component. Limited spatial and temporal resolution of the input data and simplified nature of the models results in the occurrence of systematic and random deviations of the model outputs when compared to the ground observations. The deviations in the satellite-computed data, which have systematic nature, can be reduced by site adaptation or regional adaptation methods. The terminology related to the procedure improving accuracy of the satellite data is not harmonized, and various terms are used: • Correlation of ground measurements and satellite-based data; • Calibration of the satellite model (its inputs and parameters); • Site adaptation or regional adaptation of satellite based data. The term site adaptation is more general and best explains the concept of adapting the satellite-based model (by correlation, calibration, fitting and recalculation) to the ground measured data. Site adaptation aims to adapt the characteristics of the satellite-based time series to the site-specific conditions described by local measurements. Three conditions are important for successful adaptation of the satellite-based model: 1. High quality DNI and GHI ground measurements for at least 12 months must be available; optimally data for 2 or 3 years should be used; 2. High quality satellite data must be used, with consistent quality over the whole period of data; 3. There must be identified a systematic difference between both data sources. Systematic difference can be measured by two characteristics: • Bias (offset) • Systematic deviation in the distribution of hourly or daily values (in the histogram) © 2017 Solargis page 25 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Systematic difference can be stable over the year or it can slightly change seasonally for certain meteorological conditions (e.g. typical cloud formation during a day, seasonal air pollution). The data analysis should distinguish systematic differences from those arising at occasional events, such as extreme sand storms or forest fires. The episodically-occurring differences may mislead the results of adaptation, especially if short period of ground measurements is only available. If one of the three above-mentioned conditions is not fulfilled, site adaptation (regional adaptation) will not provide the expected results. On the contrary, such an attempt may provide worse results. For the quantitative assessment of the accuracy enhancement procedures, the following metrics is used: • Metrics based on the comparison of all pairs of the hourly daytime data values: Mean Bias, and Root Mean Square Deviation (RMSD), histogram, in an absolute and relative form (divided by the daytime mean DNI values); • Metrics based on the difference of the cumulative distribution functions: KSI (Kolmogorov-Smirnov test Integral) [19] The normalized KSI is defined as an integral of absolute differences of two cumulative distribution functions D normalized by the integral of critical value acritical: , where critical value depends on the number of the data pairs N. As the KSI value is dependent on the size of the sample, the KSI measure may be used only for the relative comparison of fit of cumulative distribution of irradiance values. For the accuracy enhancement of solar resource parameters in this study, a combination of two methods was used. First, systematic deviations due to influence of aerosols were partially removed. Afterwards, to improve the distribution of values, the fitting of cumulative frequency distribution curves of ground measurements and satellite data was used. The site-adaptation procedure first identifies the sources of discrepancies by comparing the ground-measured data with Solargis model data, for the period of the overlap between both data sets. Based on this analysis, correction coefficients to improve the fit between the measured and the model Solargis data are developed. In the second step, these coefficients are used for the adaptation of the full (18 years) time series. The satellite data is available in 30-minute time step and the ground measurements in 1-minute time step. To partially remove the conceptual difference of point and satellite pixel measurements, prior to site adaptation, all the measures are calculated using aggregated data in the hourly time step. The adaptation was based on measured DNI data from RSP 4G instrument and GHI data from the secondary standard SR20-T1 instrument. GHI measured by the RSP 4G was not used because of higher uncertainty of the outputs (Chapter 3.3) 4.3 Results of the model adaptation at four sites The original Solargis data show a regional pattern of slight overestimation, compared to the ground measurements – mainly for DNI. The GHI fits the ground measurements very well. The model adaptation allowed removing large part of mismatch between satellite-based data and ground measurements. Tables 4.2 to 4.5 summarize validation of the site-adaptation results for all solar measuring stations. © 2017 Solargis page 26 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Table 4.2 Direct Normal Irradiance: bias and KSI before and after model site-adaptation Meteo station Original DNI data DNI after regional adaptation Bias Bias KSI Bias Bias KSI 2 2 [kWh/m ] [%] [-] [kWh/m ] [%] [-] Hanimaadhoo 10 2.8 61 0 0.1 45 Hulhulé 22 6.0 78 0 0.0 63 Kadhdhoo 19 5.1 83 0 0.0 71 Gan 21 5.3 92 0 0.0 83 Mean 18 4.8 79 0 0.0 66 Standard deviation 5.5 1.4 0.1 0.0 Table 4.3 Global Horizontal Irradiance: bias and KSI before and after model site-adaptation Meteo station Original GHI data GHI after regional adaptation Bias Bias KSI Bias Bias KSI 2 2 [kWh/m ] [%] [-] [kWh/m ] [%] [-] Hanimaadhoo 3 0.6 29 0 0.0 26 Hulhulé -1 -0.1 31 0 0.0 29 Kadhdhoo 0 0.0 30 0 0.0 29 Gan 3 0.7 41 0 0.0 32 Mean 1 0.3 33 0 0.0 29 Standard deviation 2.1 0.4 0.0 0.0 Table 4.4 Direct Normal Irradiance: RMSD before and after model site-adaptation Meteo station RMSD of original DNI data RMSD of DNI after regional adaptation Hourly Daily Monthly Hourly Daily Monthly [%] [%] [%] [%] [%] [%] Hanimaadhoo 31.4 17.0 6.2 31.4 16.8 5.4 Hulhulé 35.5 19.7 7.4 34.9 19.0 4.8 Kadhdhoo 36.4 18.6 6.4 36.0 18.0 3.8 Gan 33.8 17.3 6.9 33.4 16.7 4.4 Mean 34.3 18.1 6.7 33.9 17.6 4.6 Table 4.5 Global Horizontal Irradiance: RMSD before and after model site-adaptation Meteo station RMSD of original GHI data RMSD of GHI after regional adaptation Hourly Daily Monthly Hourly Daily Monthly [%] [%] [%] [%] [%] [%] Hanimaadhoo 15.2 6.8 2.4 15.2 6.8 2.4 Hulhulé 16.2 7.0 2.1 16.3 7.0 2.2 Kadhdhoo 16.8 6.7 1.5 16.8 6.7 1.5 Gan 15.1 6.3 2.0 15.1 6.2 1.9 Mean 15.9 6.7 2.0 15.9 6.7 2.0 As a result, at the level of individual measurement sites in Maldives, the mean bias of the adapted values was reduced to zero. The values of RMSD and KSI accuracy parameters are also reduced, especially for DNI. © 2017 Solargis page 27 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 The effect of the site adaptation is presented in a detail for all sites (Figures 4.1 to 4.4). The changes are very small as the original data have good fit to ground measurements. Hanimaadhoo: Original DNI Hanimaadhoo: DNI after adaptation Hanimaadhoo: Original GHI Hanimaadhoo: GHI after adaptation Figure 4.1: Correction of DNI and GHI hourly values for Hanimaadhoo. Left: original Solargis data, right: site-adapted Solargis data. The X-axis represents the measured data and the Y-axis represents the satellite-derived data. © 2017 Solargis page 28 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Hulhulé: Original DNI Hulhulé: DNI after adaptation Hulhulé: Original DNI Hulhulé: DNI after adaptation Figure 4.2: Correction of DNI and GHI hourly values for Hulhulé Left: original Solargis data, right: site-adapted Solargis data. The X-axis represents the measured data and the Y-axis represents the satellite-derived data. © 2017 Solargis page 29 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Kadhdhoo: Original DNI Kadhdhoo: DNI after adaptation Kadhdhoo: Original DNI Kadhdhoo: DNI after adaptation Figure 4.3: Correction of DNI and GHI hourly values for Kadhdhoo. Left: original Solargis data, right: site-adapted Solargis data. The X-axis represents the measured data and the Y-axis represents the satellite-derived data. © 2017 Solargis page 30 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Gan: Original DNI Gan: DNI after adaptation Gan: Original DNI Gan: DNI after adaptation Figure 4.4: Correction of DNI and GHI hourly values for Gan. Left: original Solargis data, right: site-adapted Solargis data. The X-axis represents the measured data and the Y-axis represents the satellite-derived data. The change of model GHI and DNI after adaptation is presented on an example of Hanimaadhoo (Figure 4.5). The change for GHI is negligible, the adapted DNI is slightly lower than original one. The other sites are very similar (Table 4.6). The site-adapted model values better represent the geographical variability of DNI and GHI solar resource and they also improve the distribution and match of hourly values. The measurements show that the model performs well in the region, and these results significantly improve the confidence about the reliability of the measured and modelled solar resource data for Maldives. © 2017 Solargis page 31 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 4.5: Comparison of Solargis original and site-adapted data for Hanimaadhoo site. Left: DNI; Right: GHI; Data represent years 1999 to 2016. Table 4.6 Comparison of long term average of yearly summaries of original and site-adapted values Meteo station DNI annual values* GHI annual values* Original Site-adapted Difference Original Site-adapted Difference 2 2 2 2 [kWh/m ] [kWh/m ] [%] [kWh/m ] [kWh/m ] [%] Hanimaadhoo 1521 1478 -2.8 2029 2020 -0.5 Hulhulé 1614 1520 -5.8 2048 2053 +0.2 Kadhdhoo 1678 1593 -5.1 2054 2056 +0.1 Gan 1742 1652 -5.1 2073 2061 -0.6 © 2017 Solargis page 32 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 5 METEOROLOGICAL MODEL DATA 5.1 Meteorological model For the territory of Maldives, the last 18 years of the Solargis model-based meteorological data is derived from the regional models. The meteorological data in Solargis database is derived from the two data sources: CFSR and CFSv2, with original characteristics specified in Table 5.1. Table 5.1 Original source of Solargis meteorological data: models CFSR and CFSv2. Climate Forecast System Climate Forecast System Reanalysis (CFSR) (CFSv2) Time period 1999 to 2010 2011 to the present time Original spatial resolution 30 x 35 km 19 x 22 km Original time resolution 1 hour 1 hour Table 5.2 shows meteorological parameters available in Solargis, their specifications and it also indicates, which of them have been delivered within this study. In general, the original spatial resolution of the models is enhanced to 1 km for air temperature and relative humidity by spatial disaggregation and use of the Digital Elevation Model SRTM-3, although for Maldives this correction has no effect. The spatial resolution of other parameters is unchanged. Table 5.2 Solargis meteorological parameters delivered within this project Time Spatial Data Data Meteorological parameter Acronym Unit resolution representation delivered validated Air temperature at 2 metres TEMP °C 60 minute 1 km Yes Yes (dry bulb temperature) Relative humidity at 2 metres RH % 60 minute 1 km Yes Yes 2 Wind speed at 10 metres WS m/s 60 minute Original model Yes Yes Wind direction at 10 metres WD ° 60 minute Original model Yes Yes Atmospheric pressure AP hPa 60 minute Original model Yes Yes Precipitable water PWAT 60 minute Original model Yes No Important note: meteorological parameters are derived from the numerical weather model outputs and they have lower spatial and temporal resolution. Thus, they do not represent the same accuracy as the solar resource data. Especially wind speed data has higher uncertainty, and it provides only overview information for solar energy projects. Thus, the local microclimate of the site may deviate from the values derived from the Solargis global database. 5.2 Validation of meteorological data The validation procedure was carried out to compare the modelled data with ground-measurements from the 4 meteorological stations installed within the ESMAP project: Hanimaadhoo, Hulhulé, Kadhdhoo and Gan. In general, the data from the meteorological models represent larger area, and it is not capable to represent accurately the local microclimate created by small land mass of the islands in the atolls of the Maldives. © 2017 Solargis page 33 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 5.2.1 Air temperature at 2 metres Air temperature is derived from the model outputs of both CSFR and CSFv2 meteorological models and recalculated at the spatial resolution of 1 km (Table 5.3 and Figures 5.1 to 5.4). Considering spatial and time interpolation, the deviation of the modelled values to the ground observations for hourly values can occasionally reach several degrees of Celsius. Figures 5.1 to 5.4 show graphical representation of the model values accuracy at the meteorological stations. In general, the model matches the ground measurements well. Main issue identified is underestimation of daily temperature amplitude. This is caused by relatively small land mass of the islands (in comparison to the pixel size of the meteorological model). Model air temperature is driven mainly by the air temperature over the ocean, where the daily amplitude is lower than temperature amplitude at individual islands. Yet, the insufficiency of meteorological models ay have limited importance, as air temperature is changing only a little across the seasons and day time. Table 5.3 Air temperature at 2 m: accuracy indicators of the model outputs [ºC]. CFSv2 model Meteorological station Bias Bias Bias Bias Bias RMSD RMSD RMSD mean min max nigh-time day-time hourly daily monthly Hanimaadhoo -0.7 0.9 -2.3 0.1 -1.6 1.6 0.9 0.8 Hulhulé -0.9 0.4 -2.2 -0.2 -1.5 1.5 1.1 0.9 Kadhdhoo -0.4 1.3 -2.2 0.5 -1.3 1.6 0.8 0.5 Gan -0.4 1.2 -1.9 0.4 -1.2 1.5 0.8 0.5 Figure 5.1: Scatterplots of air temperature at 2 m at Hanimaadhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis) Blue: day-time, Black: night-time measurements © 2017 Solargis page 34 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.2: Scatterplots of air temperature at 2 m at Hulhulé meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis) Blue: day-time, Black: night-time measurements Figure 5.3: Scatterplots of air temperature at 2 m at Kadhdhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis) Blue: day-time, Black: night-time measurements © 2017 Solargis page 35 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.4: Scatterplots of air temperature at 2 m at Gan meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis) Blue: day-time, Black: night-time measurements 5.2.2 Relative humidity Relative humidity is calculated from the specific humidity, air pressure and the site-adapted air temperature. The comparison of the model values with ground measurements at all 4 meteorological stations is shown in Table 5.4 and Figures 5.5 to 5.8. The issue identified is amplitude reduction of daily relative humidity. Yet, this effect is of limited importance, as relative humidity is quite stable across the seasons and day time. Table 5.4 Relative humidity: accuracy indicators of the model outputs [%]. CFSv2 model Meteorological station Bias Bias Bias Bias Bias RMSD RMSD RMSD mean min max nigh-time day-time hourly daily monthly Hanimaadhoo -7 -2 -12 -10 -4 9 7 7 Hulhulé -5 0 -9 -7 -2 7 5 5 Kadhdhoo -7 -3 -12 -10 -5 9 8 7 Gan -7 -3 -13 -10 -4 9 8 7 © 2017 Solargis page 36 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.5: Scatterplots of relative humidity at 2 m at Hanimaadhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time measurements. Figure 5.6: Scatterplots of relative humidity at 2 m at Hulhulé meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time measurements. © 2017 Solargis page 37 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.7: Scatterplots of relative humidity at 2 m at Kadhdhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time measurements. Figure 5.8: Scatterplots of relative humidity at 2 m at Gan meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time measurements. © 2017 Solargis page 38 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 5.2.3 Wind speed and wind direction at 10 metres Wind speed and direction values delivered within Solargis data represent the height at 10 meters and they are calculated from the CFSR and CFSv2 models, from 10 m wind u- and v- components. The spatial resolution is kept as in the original data. The height of wind measurement is not given. Wind measurements take place at the height of 3 metres, while the model data represent values at height of 10 m above ground. Comparison of the modelled wind speed with ground measurements is shown in Table 5.5 and Figures 5.9 to 5.12. The model values underestimate the wind conditions measured at the meteorological stations. Similarly to relative humidity, the data representation for wind speed and wind direction strongly depends on the local conditions; therefore the model values are only indicative and better characterize a larger region rather than the local microclimate. The important source of systematic difference is different height of the installed sensor (3 metres above ground), compared to the model assumptions (10 metres). Table 5.5 Wind speed: accuracy indicators of the model outputs [m/s]. CFSR and CFSv2 models Meteorological station Bias Bias Bias Bias Bias RMSD RMSD RMSD mean min max nigh-time day-time hourly daily monthly Hanimaadhoo 2.9 2.4 3.2 3.1 2.6 3.2 3.1 3.2 Hulhulé 2.3 2.2 2.4 2.5 2.2 2.8 2.6 2.6 Kadhdhoo 3.0 2.5 3.2 3.1 2.8 3.4 3.3 3.1 Gan 1.4 1.5 1.4 1.8 1.1 2.1 1.9 1.5 Figure 5.9: Scatterplots of wind speed at Hanimaadhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time. (observations at 3 m heigth and model data at 10 m) © 2017 Solargis page 39 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.10: Scatterplots of wind speed at Hulhulé meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time. (observations at 3 m heigth and model data at 10 m) Figure 5.11: Scatterplots of wind speed at Kadhdhoo meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time. (observations at 3 m heigth and model data at 10 m) © 2017 Solargis page 40 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 5.12: Scatterplots of wind speed at Gan meteorological station. Measured values (horizontal axis) and meteorological model values (vertical axis). Blue: day-time, black: night-time. (observations at 3 m heigth and model data at 10 m) 5.3 Uncertainty of meteorological model data The meteorological parameters are derived from two very similar numerical meteorological models covering periods from 1999 to 2010 (CFSR model) and 2011 to 2016 (CFSv2). Considering the comparison results, the uncertainty of the estimate for the main meteorological parameters is summarised in Table 5.6. It was found that the modelled air temperature fits reasonably well the measured data though (logically) due to the spatial resolution, the model daily amplitude is significantly smaller than amplitude of measured data. Similarly to air temperature, relative humidity from the model exhibits much smaller daily amplitude than measured relative humidity. Wind speed data, obtained from the meteorological model, represents an area of larger region and this data is smoothed, in comparison to the point measurements collected at the meteorological sites. For all stations at Maldives wind speed is overestimated by the meteorological model. The reason is coarse resolution of the model and different elevation (observations at 3 m heigth and model data at 10 m) Table 5.6 Expected uncertainty of modelled meteorological parameters at the project site. Unit Annual Monthly Hourly Air temperature at 2 m °C ±1.0 ±1.0 ±2.0 Relative Humidity at 2 m % ±8 ±8 ±15 Average wind speed at 10 m m/s +2.5 ±3.0 ±3.5 © 2017 Solargis page 41 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 6 SOLAR RESOURCE: UNCERTAINTY OF LONGTERM ESTIMATES 6.1 Uncertainty of solar resource yearly estimate The uncertainty of site-adapted satellite-based GHI and DNI is determined by the uncertainty of the model and of the ground measurements [7], more specifically it depends on: 1. Parameterization and adaptation of numerical models integrated in Solargis for the given data inputs and their ability to generate accurate results for various geographical and time-variable conditions: • Data inputs into Solargis model (accuracy of satellite data, aerosols and water vapour and Digital Terrain Model). • Clear-sky model and its capability to properly characterize various states of the atmosphere • Simulation accuracy of the satellite model and cloud transmittance algorithms, being able to properly distinguish different types of desert surface, clouds, fog, but also snow and ice. • Diffuse and direct decomposition models • Site adaptation methods. 2. Uncertainty of the ground-measurements, which is determined by: • Accuracy of the instruments • Maintenance practices, including sensor cleaning, calibration • Data post-processing and quality control procedures. Accuracy statistics, such as bias and RMSD (Chapter 4.3) characterize accuracy of Solargis model in the given validation points, relative to the ground measurements. The validation statistics is affected by local geography and by quality and reliability of the ground-measured data. Therefore the validation statistics only indicates performance of the model in the region. Solargis model uncertainty is compared to the high-quality data measured by the meteorological instruments. Representativeness of such data comparison (satellite and ground-measured) is determined by the precision of the measuring instruments, the maintenance and operational practices, and by quality control of the measured data – in other words, by the measurement accuracy achieved at each meteorological station. From the user’s perspective, the information about the model uncertainty has probabilistic nature. It generalizes the validation accuracy and it has to be considered at different confidence levels. The expert estimate of the calculation uncertainty in this report assumes 80% probability of occurrence of values. The solar model uncertainty is discussed in Chapters 4 and 6.1. The main findings are summarized in Table 6.1. The site-adaptation procedure reduced uncertainty of estimate of all parameters. Chapter 6.3 evaluates combined uncertainty, in which also interannual variability is included (Chapter 6.2). Even though the physical reductionfo themodel uncertainty is seen to vbe low, important component of the uncertainty reduction in Table 6.1 is dramatically increased confidence in the data values. Table 6.1 Uncertainty of the model estimates for original and site-adapted annual long-term values (Considers 80% probability of occurrence) Uncertainty of longterm Acronym Uncertainty of the original Uncertainty of the Solargis model annual values Solargis model after site adaptation Global Horizontal Irradiation GHI ±5.0% ±3.5% Direct Normal Irradiance DNI ±11.0% ±6.0% © 2017 Solargis page 42 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 6.2 Uncertainty due to interannual variability of solar radiation Weather changes in cycles and has also stochastic nature. Therefore annual solar radiation in each year can deviate from the long-term average in the range of few percent. The estimation of the interannual variability below shows the magnitude of this change. The uncertainty of GHI and DNI prediction is highest if only one single year is considered, but when averaged for a longer period, weather oscillations even out and approximate to the long-term average. In this report, the interannual variability is calculated from the unbiased standard deviation stdev of GHI and DNI over 18 years, considering, in the long-term, the normal distribution of the annual sums for n years, where xi is any particular year and is longterm yearly average. Due to the limited number of years of available data, for the calculation we apply simplified assumption of normal distribution of yearly values: 5 3 = ?45(? − )2 3=5 Tables 6.2 and 6.3 show GHI and DNI values that are to be exceeded at P90 for a consecutive number of years. The variability (varn) for a number of years (n) is calculated from the unbiased standard deviation (stdev): +C.)D 3 = 3 The uncertainty, which characterises 80% probability of occurrence (Uvar), is calculated from the variability (varn), multiplying it with 1.28155: = 1.28155 The lower boundary (negative value) of uncertainty represents 90% probability of exceedance, and it is used for calculating the P90 value. Table 6.2 Annual GHI that should be exceeded with 90% probability in the period of 1 to 10 (25) years Hanimaadhoo │Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 1.3 0.9 0.8 0.7 0.6 0.5 0.5 0.5 0.4 0.4 0.3 Uncertainty P90 [±%] 1.7 1.2 1.0 0.9 0.8 0.7 0.6 0.6 0.6 0.5 0.3 Minimum GHI P90 1985 1995 2000 2002 2004 2005 2007 2007 2008 2009 2013 Hulhulé│Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 1.5 1.0 0.9 0.7 0.7 0.6 0.6 0.5 0.5 0.5 0.3 Uncertainty P90 [±%] 1.9 1.3 1.1 1.0 0.9 0.8 0.7 0.7 0.6 0.6 0.4 Minimum GHI P90 2014 2025 2030 2033 2035 2037 2038 2039 2040 2040 2045 Kadhdhoo│Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 2.0 1.4 1.1 1.0 0.9 0.8 0.7 0.7 0.7 0.6 0.4 Uncertainty P90 [±%] 2.5 1.8 1.5 1.3 1.1 1.0 1.0 0.9 0.8 0.8 0.5 Minimum GHI P90 2004 2019 2026 2030 2032 2034 2036 2037 2038 2039 2045 Gan │Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 2.0 1.4 1.2 1.0 0.9 0.8 0.8 0.7 0.7 0.6 0.4 Uncertainty P90 [±%] 2.6 1.8 1.5 1.3 1.2 1.1 1.0 0.9 0.9 0.8 0.5 Minimum GHI P90 2008 2024 2031 2035 2038 2040 2041 2043 2044 2045 2051 © 2017 Solargis page 43 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Table 6.3 Annual DNI that should be exceeded with 90% probability in the period of 1 to 10 (25) years. Hanimaadhoo │Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 3.6 2.5 2.1 1.8 1.6 1.5 1.3 1.3 1.2 1.1 0.7 Uncertainty P90 [±%] 4.6 3.2 2.6 2.3 2.0 1.9 1.7 1.6 1.5 1.4 0.9 Minimum DNI P90 1411 1431 1439 1445 1448 1451 1453 1455 1456 1457 1465 Hulhulé │Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 3.9 2.8 2.3 2.0 1.7 1.6 1.5 1.4 1.3 1.2 0.8 Uncertainty P90 [±%] 5.0 3.5 2.9 2.5 2.2 2.0 1.9 1.8 1.7 1.6 1.0 Minimum DNI P90 1444 1466 1476 1482 1486 1489 1491 1493 1494 1496 1505 Kadhdhoo│Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 4.1 2.9 2.4 2.1 1.9 1.7 1.6 1.5 1.4 1.3 0.8 Uncertainty P90 [±%] 5.3 3.8 3.1 2.7 2.4 2.2 2.0 1.9 1.8 1.7 1.1 Minimum DNI P90 1508 1533 1544 1550 1555 1558 1561 1563 1564 1566 1576 Gan │Years 1 2 3 4 5 6 7 8 9 10 25 Variability [±%] 4.1 2.9 2.3 2.0 1.8 1.7 1.5 1.4 1.4 1.3 0.8 Uncertainty P90 [±%] 5.2 3.7 3.0 2.6 2.3 2.1 2.0 1.8 1.7 1.6 1.0 Minimum DNI P90 1566 1591 1603 1609 1614 1617 1620 1622 1624 1625 1635 We can interpret the above Table 6.2 and 6.3 on the example of Hulhulé site: i. GHI interannual variability at P90 of 1.9% has to be considered for any single year in Hulhule. In other words, 2 assuming that the long-term average is 2053 kWh/m , it is expected (with 90% probability) that annual GHI 2 exceeds, at any single year, the value of 2014 kWh/m . ii. Within a period of three consecutive years, it is expected at P90 that annual average of GHI exceeds value 2 of 2030 kWh/m ; iii. For a period of 25 years, it is expected at 90% probability that due to interannual variability the estimate of the long-term annual DNI average will deviate within the range of ±1.0% in Hulhulé. Thus assuming that the 2 estimate of the long-term average is 1520 kWh/m , it can be expected at P90 that due to variability of 2 weather, it should be at least 1505 kWh/m . It is to be underlined that prediction of the future irradiation is based on the analysis of the recent historical data (period 1999 to 2016). Future weather changes may include man-induced or natural events such as volcano eruptions, which may have impact on this prediction. Based on the existing scientific knowledge [20, 21], an effect of extreme volcano eruptions, with an emission of large amount of stratospheric aerosols, can be estimated on the example of Pinatubo event in 1991 (the second th largest volcano eruption in 20 century). It can be expected that in such a case, the annual DNI in the affected year can decrease by approx. 16% or more, compared to the long-term average, still influencing another two consecutive years. In the same way, the volcano eruption of the comparable size may reduce long-term average estimate of DNI by about 4%. The decrease of GHI is much lower; the annual value in the particular year of eruption could be reduced by about 2% compared to the long-term average. 6.3 Combined uncertainty In this Chapter, the combined uncertainty of the annual GHI and DNI values is quantified. Taking into account uncertainties of both types of data (satellite and ground measured), the combined effect of two components of the uncertainty of the site-adapted GHI and DNI values has to be considered. 1. Uncertainty of the estimate (Uest) of the annual solar resource values, which is ±3.5% for GHI and ±6.0% for DNI (Chapter 6.1); 2. Interannual variability (Uvar) in any particular year, due to changing weather. In four Maldivian sites, it varies from ±1.7% to ±2.6% for GHI and from ±4.6% to ±5.3% for DNI. The uncertainty due to weather variability decreases over the time with square root of the number of years (Chapter 6.2). The two above-mentioned uncertainties combine in Uc (see Glossary), which represents a conservative expectation of the minimum GHI and DNI assuming various number of years N (Tables 6.4 and 6.5). Considering © 2017 Solargis page 44 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 a simplified assumption of normal distribution of the annual values, probability of exceedance can be calculated at different confidence levels. GHI and DNI minimum annual values expected for combined uncertainty in any single year are shown on Figure 6.2 and 6.3. Table 6.4 Combined probability of exceedance of annual GHI for uncertainty of the estimate ±3.5%. Nr. of Uncertainty Interanual Combined Expected minimum │Hanimaadhoo 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 3.5 1.7 3.9 2162 2120 2098 2061 2020 1978 1941 1919 1877 5 3.5 0.8 3.6 2151 2112 2092 2058 2020 1981 1947 1927 1888 10 3.5 0.5 3.5 2149 2111 2091 2057 2020 1982 1948 1928 1890 25 3.5 0.3 3.5 2148 2111 2091 2057 2020 1982 1949 1928 1891 Nr. of Uncertainty Interanual Combined Expected minimum│Hulhulé 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 3.5 1.9 4.0 2201 2158 2134 2096 2053 2010 1971 1948 1904 5 3.5 0.9 3.6 2187 2148 2127 2092 2053 2014 1979 1958 1918 10 3.5 0.6 3.6 2185 2146 2126 2091 2053 2014 1980 1959 1920 25 3.5 0.4 3.5 2184 2145 2125 2091 2053 2015 1980 1960 1921 Nr. of Uncertainty Interanual Combined Expected minimum│Kadhdhoo 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 3.5 2.5 4.3 2217 2170 2144 2102 2056 2009 1967 1942 1895 5 3.5 1.1 3.7 2193 2153 2131 2095 2056 2016 1980 1959 1918 10 3.5 0.8 3.6 2190 2150 2129 2095 2056 2017 1982 1961 1922 25 3.5 0.5 3.5 2188 2149 2128 2094 2056 2017 1983 1962 1924 Nr. of Uncertainty Interanual Combined Expected minimum │Gan 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 3.5 2.6 4.4 2224 2177 2151 2109 2061 2014 1972 1946 1898 5 3.5 1.2 3.7 2199 2159 2137 2101 2061 2021 1985 1964 1923 10 3.5 0.8 3.6 2196 2157 2136 2100 2061 2022 1987 1966 1927 25 3.5 0.5 3.5 2194 2155 2134 2100 2061 2023 1989 1968 1929 Table 6.5 Combined probability of exceedance of annual DNI for uncertainty of the estimate ±6.0%. Nr. of Uncertainty Interanual Combined Expected minimum│Hanimaadhoo 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 6.0 4.6 7.5 1681 1621 1590 1537 1478 1420 1367 1335 1276 5 6.0 2.0 6.3 1649 1599 1572 1528 1478 1429 1385 1358 1308 10 6.0 1.4 6.2 1644 1596 1570 1526 1478 1430 1387 1361 1313 25 6.0 0.9 6.1 1641 1594 1568 1526 1478 1431 1389 1363 1316 Nr. of Uncertainty Interanual Combined Expected minimum│Hulhulé 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 6.0 5.0 7.8 1735 1672 1638 1582 1520 1457 1401 1367 1304 5 6.0 2.2 6.4 1696 1645 1617 1571 1520 1469 1422 1395 1343 10 6.0 1.6 6.2 1691 1641 1614 1569 1520 1470 1425 1399 1349 25 6.0 1.0 6.1 1688 1638 1612 1568 1520 1471 1427 1401 1352 © 2017 Solargis page 45 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Nr. of Uncertainty Interanual Combined Expected minimum │Kadhdhoo 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 6.0 5.3 8.0 1824 1756 1720 1660 1593 1525 1465 1429 1361 5 6.0 2.4 6.5 1779 1725 1695 1647 1593 1539 1490 1461 1406 10 6.0 1.7 6.2 1773 1720 1692 1645 1593 1540 1493 1465 1412 25 6.0 1.1 6.1 1769 1717 1690 1644 1593 1542 1496 1468 1416 Nr. of Uncertainty Interanual Combined Expected minimum │Gan 2 years of estimate variability uncertainty [kWh/m ] N [±%] N years [±%] P90 [±%] P01 P05 P10 P25 P50 P75 P90 P95 P99 1 6.0 5.2 7.9 1890 1821 1783 1721 1652 1583 1521 1484 1414 5 6.0 2.3 6.4 1845 1789 1759 1708 1652 1596 1546 1516 1459 10 6.0 1.6 6.2 1839 1784 1755 1706 1652 1598 1549 1520 1466 25 6.0 1.0 6.1 1835 1781 1753 1705 1652 1599 1552 1523 1470 This analysis is based on the data representing a history of year 1999 to 2016, and on the expert extrapolation of the related weather variability. This report may not reflect possible man-induced climate change or occurrence of extreme events such as large volcano eruptions in the future (see the last paragraph in Chapter 6.2). Graphical visualisation of Tables 6.4 and 6.5 on the example of Hulhulé Airport is shown in Figures 6.1 and 6.2, where the expected probabilities of exceedance (different Pxx scenarios) are drawn on the cumulatove distribution curve showing yearly GHI amd DNI values. 100 P99: 1904 P95: 1948 GHI Value at Pxx P50 P90: 1971 P75 P90 90 P95 P99 80 P75: 2010 70 60 P50: 2053 Pxx 50 40 30 20 10 0 1800 1850 1900 1950 2000 2050 2100 2150 2200 2250 GHI [kWh/m2] Figure 6.1: Expected Pxx values for GHI at Hulhulé Airport © 2017 Solargis page 46 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 100 P99: 1304 P95: 1367 DNI Value at Pxx P50 P90: 1401 P75 P90 90 P95 P99 80 P75: 1457 70 60 P50: 1520 Pxx 50 40 30 20 10 0 1200 1300 1400 1500 1600 1700 1800 DNI [kWh/m2] Figure 6.2: Expected Pxx values for DNI at Hulhulé Airport © 2017 Solargis page 47 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 7 TIME SERIES AND TYPICAL METEOROLOGICAL YEAR DATA 7.1 Delivered data sets Table 7.1 Parameters in the delivered site-adapted time series and TMY data (hourly time step) Parameter Acronym Unit TS TMY P50 TMY P90 2 Global horizontal irradiance GHI W/m X X X 2 Direct normal irradiance DNI W/m X X X 2 Diffuse horizontal irradiance DIF W/m X X X 2 Global tilted irradiance (at optimum angle) GHI W/m X - - Solar azimuth SA ° X X X Solar elevation SE ° X X X Air temperature at 2 metres TEMP °C X X X Wind speed at 10 metres WS m/s X X X Wind direction at 10 metres WD ° X X X Relative humidity RH % X X X Air Pressure AP hPa X X X 2 Precipitable Water PWAT kg/m X X X 7.2 TMY method The Typical Meteorological Year (TMY) data sets are delivered, together with Solargis time series data and this report. TMY contains hourly data derived from the time series covering complete 18 years (1999 to 2016). The data history of 18 years is compressed into one year (Figure 7.1 to 7.3) following two criteria: • Minimum difference between statistical characteristics (annual average, monthly averages) of TMY and long-term time series. This criterion is given about 80% weighting. • Maximum similarity of monthly Cumulative Distribution Functions (CDF) of TMY and full-time series, so that occurrence of typical hourly values is well represented for each month. This criterion is given about 20% weighting. TMY P50 data set is constructed on the monthly basis. For each month the long-term average monthly value and cumulative distribution for each parameters is calculated: Direct Normal Irradiance (DNI), Global Horizontal Irradiance (GHI), Diffuse Horizontal Irradiance (DIF) and Air Temperature (TEMP). Following the monthly data for each individual year from the set of 18 years are compared to the long-term parameters. The monthly data from the year, which resembles the long-term parameters more closely, is selected. The procedure is repeated for all 12 months, and the TMY is constructed by concatenating the selected months into one artificial (but representative) year. The method for calculation P90 data set is based on the TMY P50 method. It has been modified in a way of how a candidate month is selected. The search for set of twelve candidates is repeated in iteration until a condition of minimization of difference between annual P90 value and annual average of new TMY is reached (instead of minimization of differences in monthly means and CDFs, as applied in P50 case). Once the selection converges to minimum difference, the TMY is created by concatenation of selected months. The P90 annual values are calculated for each confidence limit − from the combined uncertainty of estimate and inter-annual variability, which can occur in any year (Chapter 6.3). © 2017 Solargis page 48 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 To derive TMY that fits specific needs of the selected energy application the different weights are given to individual parameters – thus highlighting important properties. In solar energy applications, the higher importance is given to GHI and DNI. In assembling TMY P50, the values of DNI, GHI, DIF and TEMP are only considered, where the weights are set as follows: 0.9 is given to DNI, 0.3 to GHI, 0.02 to diffuse horizontal irradiance, and 0.07 to air temperature (divided by the total of 1.29). To derive solar resource parameters with an hourly time step, the original satellite data with time resolution of 30-minutes were aggregated by time integration. The meteorological parameters are available in the original 1-hourly time step. The TMY datasets were constructed from solar radiation and meteorological data (Chapters 4 and 5). Time zone was adjusted to Maldives Time MVT (UTC +05:00). More about the Solargis TMY method in [22]. 7.3 Results Two data sets are derived from the Solargis historical time series for the four sites: P50 and P90. In graphts and tables below we show the values for Hulhulé meteorological site, in order to present the methodology of TMY data calculation. Important note: Due to the inherent features of the underlying methods, monthly values in TMY data set in Tables 7.2 to 7.4 do not fit to the values generated from full time series (Figures 7.1 to 7.3). Table 7.2 Monthly and yearly long-term GHI averages as calculated from time series and from TMY representing P50, and P90 cases at Hulhulé site Global Horizontal 2 1 2 3 4 5 6 7 8 9 10 11 12 Year Irradiation [kWh/m ] Time series (18 years) 178 182 208 184 167 157 161 169 163 176 153 155 2053 TMY for P50 case 178 176 204 187 165 157 166 170 157 176 161 154 2053 TMY for P90 case 172 177 206 175 157 140 158 168 155 167 144 151 1971 Table 7.3 Monthly and yearly long-term DNI averages as calculated from time series and from TMY representing P50, and P90 cases at Hulhulé site Direct Normal 2 1 2 3 4 5 6 7 8 9 10 11 12 Year Irradiation [kWh/m ] Time series (18 years) 143 151 169 146 120 107 106 109 104 131 115 119 1520 TMY for P50 case 142 151 169 146 121 107 106 109 103 132 115 119 1520 TMY for P90 case 135 146 166 136 103 85 104 107 95 120 97 107 1401 Table 7.4 Monthly and yearly long-term TEMP averages as calculated from time series and from TMY representing P50, and P90 cases at Hulhulé site Air temperature [°C] 1 2 3 4 5 6 7 8 9 10 11 12 Year Time series (18 years) 27.5 27.7 28.2 28.7 28.7 28.4 28.2 28.0 27.9 27.9 27.7 27.5 28.0 TMY for P50 case 27.2 27.5 28.4 28.2 28.4 28.7 28.0 28.1 27.9 27.9 28.2 27.5 28.0 TMY for P90 case 27.5 27.5 28.2 28.7 28.6 28.5 28.1 27.9 27.8 28.0 27.6 27.5 28.0 As an example o finterpretation of the tables above, the TMY data sets for P50 and P90 can be described as: 1. P50 TMY data set represents, for each month, the average climate conditions and the most representative cumulative distribution function, therefore extreme situations (e.g. extremely cloudy weather) are not represented in this dataset. The long-term annual summary of GHI and DNI are considered as the most 2 critical parameters to consider, and in this data set P50 GHI value is 2053 kWh/m and DNI value is 2 1520 kWh/m . © 2017 Solargis page 49 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 2. P90 TMY data set represents for each month the climate conditions, which after summarization GHI and DNI for the whole year results in the value equal or close to P90 derived by the analysis of uncertainty of the estimate and of the interannual variability for any single year (Chapter 6.3). Thus TMY for P90 represents 2 generally a conservative estimate, i.e. a year with the long-term value of GHI of 1971 kWh/m and DNI of 2 1401 kWh/m . 300 250 GHI [kWh/m2] 200 150 100 50 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Time series TMY P50 TMY P90 Figure 7.1: GHI monthly values derived from time series and TMY P50 and P90 at Hulhulé site. 300 250 DNI [kWh/m2] 200 150 100 50 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Time series TMY P50 TMY P90 Figure 7.2: DNI monthly values derived from time series and TMY P50 and P90 at Hulhulé site. © 2017 Solargis page 50 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 30.0 29.0 TEMP [°C] 28.0 27.0 26.0 25.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Time series TMY P50 TMY P90 Figure 7.3: TEMP monthly values derived from time series and TMY P50 and P90 at Hulhulé site It is important to note that the data reduction in the TMY data set is not possible without loss of information contained in the original multiyear time series. Therefore time series data are considered as the most accurate reference suitable for the statistical analysis of solar resource and meteorological parameters of the site. Figure 7.4: Seasonal profile of GHI, DNI and DIF for Typical Meteorological Year P50 2 Hulhulé site: X-axis – day of the year; Y-axis – solar irradiance W/m © 2017 Solargis page 51 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure 7.5: Snapshot of Typical Meteorological Year for P50 for Hulhulé © 2017 Solargis page 52 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 8 CONCLUSIONS This report accompanies delivery of site-specific solar resource and meteorological data for four sites, where solar meteorological stations have been installed and operated during the year 2016. The measurement campaign is ongoing, and the same data exercise will be repeated after concluding of 24 months measurements, with expectation for reducing som more the data uncertainty. The measured data used in site-adaptation of the Solargis model. As a result, reliable historical time series and TMY data is computed and provided in formats ready to use in standard phovotoltaic energy simulation software. The reliability of the delivered data is high and it can be used for energy evaluation of any project in Maldives and also for bankable assessment. © 2017 Solargis page 53 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 ANNEX 1: SITE RELATED DATA STATISTICS Yearly summaries of solar and meteorological parameters Statistics for site adapted model yearly values representing 18 years (2999 to 2016). Average yearly sum of Global Horizontal Irradiation [kWh/m2] Average daily sum of Global Horizontal Irradiation [kWh/m2] 6.0 2192 5.5 2009 5.0 1826 4.5 1644 4.0 1461 3.5 1278 3.0 1096 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Year Hanimaadhoo 1.7% Hulhulé 1.9% Kadhdhoo 2.5% Gan 2.6% 2 Figure I: Interannual variability of site-adapted yearly GHI [kWh/m ]. Annual average (avg, solid line) and standard deviation (value behind the names of sites). 2 Figure II: Interannual variability of site-adapted yearly DNI [kWh/m ]. Annual average (avg, solid line) and standard deviation (value behind the names of sites). © 2017 Solargis page 54 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure III: Interannual variability of yearly TEMP [°C]. Annual average (avg, solid line). © 2017 Solargis page 55 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Monthly summaries of solar and meteorological parameters The graphs compare (site-adapted) monthly model time series compared to long-term averages. 8.0 8.0 7.0 Hanimaadhoo Hulhulé 7.0 Daily sums of GHI [kWh/m2] Daily sums of GHI [kWh/m2] 6.0 6.0 5.0 5.0 4.0 4.0 3.0 3.0 2.0 2.0 1.0 1.0 0.0 0.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year 8.0 8.0 7.0 Kadhdhoo Gan 7.0 Daily sums of GHI [kWh/m2] Daily sums of GHI [kWh/m2] 6.0 6.0 5.0 5.0 4.0 4.0 3.0 3.0 2.0 2.0 1.0 1.0 0.0 0.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year 2 Figure IV: GHI monthly averages [kWh/m ]. Monthly average shown as solid line; min/max monthly values as as boundary lines; last 12 months in red. 8.0 8.0 7.0 Hanimaadhoo 7.0 Hulhulé 6.0 6.0 Daily sums of DNI [kWh/m2] Daily sums of DNI [kWh/m2] 5.0 5.0 4.0 4.0 3.0 3.0 2.0 2.0 1.0 1.0 0.0 0.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year 8.0 8.0 7.0 Kadhdhoo 7.0 Gan 6.0 6.0 Daily sums of DNI [kWh/m2] Daily sums of DNI [kWh/m2] 5.0 5.0 4.0 4.0 3.0 3.0 2.0 2.0 1.0 1.0 0.0 0.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year 2 Figure V: DNI monthly averages [kWh/m ]. Monthly average shown as solid line; min/max monthly values as boundary lines; last 12 months shown in red. © 2017 Solargis page 56 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 30.0 30.0 29.0 Hanimaadhoo 29.0 Hulhulé Monthly air temperature [°C] Monthly air temperature [°C] 28.0 28.0 27.0 27.0 26.0 26.0 25.0 25.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year 30.0 30.0 29.0 Kadhdhoo 29.0 Gan Monthly air temperature [°C] Monthly air temperature [°C] 28.0 28.0 27.0 27.0 26.0 26.0 25.0 25.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec max-min LTA Last Year max-min LTA Last Year Figure VI: TEMP monthly averages [°C]. Monthly average shown as solid line; min/max onthly values as boundary lines; last 12 months shown in red. © 2017 Solargis page 57 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Frequency of occurrence of GHI and DNI daily model values for a period 1999 to 2016 The histograms below show occurrence statistics of daily values derived from the satellite-based time series for GHI and DNI. The time covered in the graphs below is 18 complete calendar years (1999 to 2016). The occurrence is calculated separately for each month. 25 12 12 25 12 25 January February March Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 25 12 12 25 April May June Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 12 25 12 25 July August September Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 25 12 12 25 October November December Percentage of days 10 10 10 20 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 5 5 2 2 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Figure VII: Histograms of daily summaries of Global Horizontal Irradiation in Hanimaadhoo. 12 25 25 12 12 25 January February March Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 12 25 12 25 25 12 April May June Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 12 25 25 12 12 25 July August September Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 12 25 25 12 October November December Percentage of days 10 10 10 20 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 5 5 2 2 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Figure VIII: Histograms of daily summaries of Global Horizontal Irradiation in Hulhulé. © 2017 Solargis page 58 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 12 25 12 25 25 12 January February March Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 25 12 12 25 April May June Percentage of days 10 10 20 10 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 12 25 12 25 25 12 July August September Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 12 25 25 12 October November December Percentage of days 10 10 10 20 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 5 5 2 2 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Figure IX: Histograms of daily summaries of Global Horizontal Irradiation in Kadhdhoo. 12 25 12 25 25 12 January February March Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 12 25 12 25 12 25 April May June Percentage of days 10 10 20 10 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 12 25 12 25 25 12 July August September Percentage of days 10 20 10 20 10 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 2 5 2 5 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 25 12 12 25 25 12 October November December Percentage of days 10 10 10 20 20 20 8 8 8 15 15 15 6 6 6 10 10 10 4 4 4 5 5 5 2 2 2 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Daily sum of GHI [kWh/m2] Figure X: Histograms of daily summaries of Global Horizontal Irradiation in Gan. Figures VII to X show histograms of daily GHI summaries for each month as calculated from Solargis time series representing the years 1999 to 2016. The distribution of daily values is not symmetric: median is drawn by the vertical line, and percentiles P10, P25, and P75, and P90 are displayed with dark grey and light grey colour bands, respectively. The percentiles P10 and P90 show 80% occurrence of daily values within each month and percentiles P25 and P75 show 50% occurrence. © 2017 Solargis page 59 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 10 10 10 January February March Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 April May June Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 July August September Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 October November December Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Figure XI: Histograms of daily summaries of Direct Normal Irradiation in Hanimaadhoo. 10 10 10 January February March Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 April May June Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 July August September Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 October November December Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Figure XII: Histograms of daily summaries of Direct Normal Irradiation in Hulhulé. © 2017 Solargis page 60 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 10 10 10 January February March Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 April May June Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 July August September Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 October November December Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Figure XIII: Histograms of daily summaries of Direct Normal Irradiation in Kadhdhoo. 10 10 10 January February March Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 April May June Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 July August September Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10 10 10 October November December Percentage of days 7.5 7.5 7.5 5 5 5 2.5 2.5 2.5 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Daily sum of DNI [kWh/m2] Figure XIV: Histograms of daily summaries of Direct Normal Irradiation in Gan. Figures XI to XIV show histograms of daily DNI summaries for each month as calculated from Solargis time series representing the years 1999 to 2016. The distribution of daily values is not symmetric: median is drawn by the vertical line, and percentiles P10, P25, and P75, and P90 are displayed with dark grey and light grey colour bands, respectively. The percentiles P10 and P90 show 80% occurrence of daily values within each month and percentiles P25 and P75 show 50% occurrence. © 2017 Solargis page 61 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Frequency of occurrence of GHI and DNI 30-minute model values for a period 1999 to 2016 The histograms below show occurrence statistics of 30-minute values derived from the satellite-based time series for GHI and DNI. The time covered in the graphs below is 18 complete calendar years (1999 to 2016). The occurrence is calculated separately for each month. Figure XV: Histograms and cumulative distribution function of 30-minute GHI in Hanimaadhoo © 2017 Solargis page 62 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XVI: Histograms and cumulative distribution function of 30-minute GHI in Hulhulé Figure XVII: Histograms and cumulative distribution function of 30-minute GHI in Kadhdhoo Figure XVIII: Histograms and cumulative distribution function of 30-minute GHI in Gan Figures XV to XVIII show monthly histograms (bars) and cumulative distribution (line) of 30-minute GHI values, 2 calculated from Solargis time series. The values represent the occurrence of GHI values within 50 W/m bins, 2 ranging from 0 to 1200 W/m . © 2017 Solargis page 63 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XIX: Histograms and cumulative distribution function of 30-minute DNI in Hanimaadhoo Figure XX: Histograms and cumulative distribution function of 30-minute DNI in Hulhulé © 2017 Solargis page 64 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XXI: Histograms and cumulative distribution function of 30-minute DNI in Kadhdhoo Figure XXII: Histograms and cumulative distribution function of 30-minute DNI in Gan Figures XIX to XXII show monthly histograms (bars) and cumulative distribution (line) of 30-minute DNI values, 2 calculated from Solargis time series. The values represent the occurrence of DNI values within 50 W/m bins, 2 ranging from 0 to 1200 W/m . © 2017 Solargis page 65 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Frequency of occurrence of GHI and DNI measured and model values representing year 2016 Figures XXIII to XXX show histograms comparing the measured values with the model GHI and DNI data. The period covered in these histogram is last 12-months (one full year of data, i.e. from 1 Jan 2016 to 31 Dec 2016): • 1-minute measured vs. 30-min satellite-based model values • 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values • Daily measured (aggregated from 1-min) vs. daily satellite-based model values Aggregation process deals with the missing values in the ground measurement in three steps: 1. Only those 1-minute measured data values that passed through quality control (Chapter 3.3) is taken into account (satellite time series does not have gaps.); 2. Aggregation of 1-minute measured data values into 30-minute slots (equivalent to satellite time slots) is applied if more than 15 valid data-points is available, otherwise the 30-minute data slot is ignored in further statistical comparison; 3. Daily aggregation of measured data represents the same 30-minute time slots in a day (passing through the two steps above), as those in the satellite-based data. Incorrect data slots found in the measurements are excluded in both the measured and model data. Figure XXIII: Measured vs. satellite-based GHI values in Hanimaadhoo 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values Figure XXIV: Measured vs. satellite-based GHI values in Hulhulé 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values © 2017 Solargis page 66 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XXV: Measured vs. satellite-based GHI values in Kadhdhoo 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values Figure XXVI: Measured vs. satellite-based GHI values in Gan 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values Figure XXVII: Measured vs. satellite-based DNI values in Hanimaadhoo 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values Figure XXVIII: Measured vs. satellite-based DNI values in Hulhulé 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values © 2017 Solargis page 67 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XXIX: Measured vs. satellite-based DNI values in Kadhdhoo 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values Figure XXX: Measured vs. satellite-based DNI values in Gan 1-minute measured vs. 30-min satellite-based values. 30-minute measured (aggregated from 1-min) vs. 30-min satellite-based values Daily measured (aggregated from 1-min) vs. daily satellite-based values © 2017 Solargis page 68 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Frequency of occurrence of GHI and DNI ramps Figures XXXI to XXXVIII show histograms of instantaneous changes (ramps) calculated from the measurements and compared to the instantaneous changes calculated fro the model data. Figures show both negative (-) and positive (+) changes. Two versions for GHI and DNI are shown: • Ramps calculated from 1-minute measured values compared to ramps calculated from 30-minute satellite-based data (figure on the left) • Ramps calculated from 30-minute aggregated valid measurement compared to ramps calculated from 30-minute satellite-based data (figure on the right). Occurrence of gaps in the measurements is managed in the same way as described about in this Chapter: 1. For measurements, only those 1-minute data values (measurements) that passed through quality control (Chapter 3.3) is taken into account (satellite time series does not have gaps.); 2. For measurements, the aggregation (averaging) of 1-minute measured data values into 30-minute slots (equivalent to satellite time slots) is applied if more than 15 valid data-points is available, otherwise the 30-minute data slot is ignored in further statistical comparison; Figure XXXI: 1-minute and 30-minute GHI ramps (measured and satellite data) at Hanimaadhoo. Figure XXXII: 1-minute and 30-minute GHI ramps (measured and satellite data) at Hulhulé © 2017 Solargis page 69 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XXXIII: 1-minute and 30-minute GHI ramps (measured and satellite data) at Kadhdhoo Figure XXXIV: 1-minute and 30-minute GHI ramps (measured and satellite data) at Gan Figure XXXV: 1-minute and 30-minute DNI ramps (measured and satellite data) at Hanimaadhoo © 2017 Solargis page 70 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure XXXVI: 1-minute and 30-minute DNI ramps (measured and satellite data) at Hulhulé Figure XXXVII: 1-minute and 30-minute DNI ramps (measured and satellite data) at Kadhdhoo Figure XXXVIII: 1-minute and 30-minute DNI ramps (measured and satellite data) at Gan © 2017 Solargis page 71 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 LIST OF FIGURES Figure 2.1: Position of solar meteorological stations in Maldives ....................................................................... 10 Figure 3.1 Results of GHI and DNI quality control in Hanimaadhoo. .................................................................... 15 Figure 3.2 Systematic difference between GHI from SR20 and RSP − Hanimaadhoo. ........................................ 16 Figure 3.3 Results of GHI and DNI quality control in Hulhulé. .............................................................................. 17 Figure 3.4 Effect of RSP shadowband issues – drop of DNI in Hulhulé. .............................................................. 18 Figure 3.5 Systematic difference between GHI from SR20 and RSP - Hulhulé. .................................................... 18 Figure 3.6 Results of GHI and DNI quality control − Kadhdhoo. ........................................................................... 19 Figure 3.7 Effect of RSP shadowband issues – drop of DNI in Kadhdhoo. .......................................................... 20 Figure 3.8 Systematic difference between GHI from SR20 and RSP − Kadhdhoo................................................ 20 Figure 3.9 Results of GHI and DNI quality control − Gan. ..................................................................................... 21 Figure 3.10 Effect of RSP shadowband issues – drop of DNI in Gan. .................................................................. 22 Figure 3.11 Systematic difference between GHI from SR20 and RSP - Gan......................................................... 22 Figure 3.12 Replacement of SR20 GHI by RSP GHI due to wind mast shading in Gan. ........................................ 22 Figure 4.1: Correction of DNI and GHI hourly values for Hanimaadhoo. .............................................................. 28 Figure 4.2: Correction of DNI and GHI hourly values for Hulhulé.......................................................................... 29 Figure 4.3: Correction of DNI and GHI hourly values for Kadhdhoo. .................................................................... 30 Figure 4.4: Correction of DNI and GHI hourly values for Gan. .............................................................................. 31 Figure 4.5: Comparison of Solargis original and site-adapted data for Hanimaadhoo site. ................................. 32 Figure 5.1: Scatterplots of air temperature at 2 m at Hanimaadhoo meteorological station. .............................. 34 Figure 5.2: Scatterplots of air temperature at 2 m at Hulhulé meteorological station. ........................................ 35 Figure 5.3: Scatterplots of air temperature at 2 m at Kadhdhoo meteorological station. .................................... 35 Figure 5.4: Scatterplots of air temperature at 2 m at Gan meteorological station. .............................................. 36 Figure 5.5: Scatterplots of relative humidity at 2 m at Hanimaadhoo meteorological station. ............................ 37 Figure 5.6: Scatterplots of relative humidity at 2 m at Hulhulé meteorological station. ....................................... 37 Figure 5.7: Scatterplots of relative humidity at 2 m at Kadhdhoo meteorological station.................................... 38 Figure 5.8: Scatterplots of relative humidity at 2 m at Gan meteorological station. ............................................ 38 Figure 5.9: Scatterplots of wind speed at Hanimaadhoo meteorological station. ............................................... 39 Figure 5.10: Scatterplots of wind speed at Hulhulé meteorological station. ........................................................ 40 Figure 5.11: Scatterplots of wind speed at Kadhdhoo meteorological station..................................................... 40 Figure 5.12: Scatterplots of wind speed at Gan meteorological station............................................................... 41 Figure 6.1: Expected Pxx values for GHI at Hulhulé Airport.................................................................................. 46 Figure 6.2: Expected Pxx values for DNI at Hulhulé Airport .................................................................................. 47 Figure 7.1: GHI monthly values derived from time series and TMY P50 and P90 ................................................ 50 Figure 7.2: DNI monthly values derived from time series and TMY P50 and P90 ................................................ 50 Figure 7.3: TEMP monthly values derived from time series and TMY P50 and P90 ............................................. 51 Figure 7.4: Seasonal profile of GHI, DNI and DIF for Typical Meteorological Year P50 ........................................ 51 Figure 7.5: Snapshot of Typical Meteorological Year for P50 for Hulhulé ............................................................ 52 Figure I: Interannual variability of site-adapted yearly GHI [kWh/m ].................................................................... 54 2 © 2017 Solargis page 72 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Figure II: Interannual variability of site-adapted yearly DNI [kWh/m ]................................................................... 54 2 Figure III: Interannual variability of yearly TEMP [°C]. ........................................................................................... 55 Figure IV: GHI monthly averages [kWh/m ]. ......................................................................................................... 56 2 Figure V: DNI monthly averages [kWh/m ]. .......................................................................................................... 56 2 Figure VI: TEMP monthly averages [°C]. ............................................................................................................... 57 Figure VII: Histograms of daily summaries of Global Horizontal Irradiation in Hanimaadhoo. ............................ 58 Figure VIII: Histograms of daily summaries of Global Horizontal Irradiation in Hulhulé. ..................................... 58 Figure IX: Histograms of daily summaries of Global Horizontal Irradiation in Kadhdhoo. ................................... 59 Figure X: Histograms of daily summaries of Global Horizontal Irradiation in Gan. .............................................. 59 Figure XI: Histograms of daily summaries of Direct Normal Irradiation in Hanimaadhoo.................................... 60 Figure XII: Histograms of daily summaries of Direct Normal Irradiation in Hulhulé. ............................................ 60 Figure XIII: Histograms of daily summaries of Direct Normal Irradiation in Kadhdhoo. ....................................... 61 Figure XIV: Histograms of daily summaries of Direct Normal Irradiation in Gan. ................................................ 61 Figure XV: Histograms and cumulative distribution function of 30-minute GHI in Hanimaadhoo........................ 62 Figure XVI: Histograms and cumulative distribution function of 30-minute GHI in Hulhulé ................................. 63 Figure XVII: Histograms and cumulative distribution function of 30-minute GHI in Kadhdhoo ............................ 63 Figure XVIII: Histograms and cumulative distribution function of 30-minute GHI in Gan..................................... 63 Figure XIX: Histograms and cumulative distribution function of 30-minute DNI in Hanimaadhoo ....................... 64 Figure XX: Histograms and cumulative distribution function of 30-minute DNI in Hulhulé .................................. 64 Figure XXI: Histograms and cumulative distribution function of 30-minute DNI in Kadhdhoo ............................. 65 Figure XXII: Histograms and cumulative distribution function of 30-minute DNI in Gan ...................................... 65 Figure XXIII: Measured vs. satellite-based GHI values in Hanimaadhoo .............................................................. 66 Figure XXIV: Measured vs. satellite-based GHI values in Hulhulé ........................................................................ 66 Figure XXV: Measured vs. satellite-based GHI values in Kadhdhoo ..................................................................... 67 Figure XXVI: Measured vs. satellite-based GHI values in Gan .............................................................................. 67 Figure XXVII: Measured vs. satellite-based DNI values in Hanimaadhoo ............................................................. 67 Figure XXVIII: Measured vs. satellite-based DNI values in Hulhulé ...................................................................... 67 Figure XXIX: Measured vs. satellite-based DNI values in Kadhdhoo .................................................................... 68 Figure XXX: Measured vs. satellite-based DNI values in Gan ............................................................................... 68 Figure XXXI: 1-minute and 30-minute GHI ramps (measured and satellite data) at Hanimaadhoo...................... 69 Figure XXXII: 1-minute and 30-minute GHI ramps (measured and satellite data) at Hulhulé ............................... 69 Figure XXXIII: 1-minute and 30-minute GHI ramps (measured and satellite data) at Kadhdhoo .......................... 70 Figure XXXIV: 1-minute and 30-minute GHI ramps (measured and satellite data) at Gan ................................... 70 Figure XXXV: 1-minute and 30-minute DNI ramps (measured and satellite data) at Hanimaadhoo .................... 70 Figure XXXVI: 1-minute and 30-minute DNI ramps (measured and satellite data) at Hulhulé .............................. 71 Figure XXXVII: 1-minute and 30-minute DNI ramps (measured and satellite data) at Kadhdhoo ........................ 71 Figure XXXVIII: 1-minute and 30-minute DNI ramps (measured and satellite data) at Gan ................................. 71 © 2017 Solargis page 73 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 LIST OF TABLES Table 1.1 Delivered data characterstics .......................................................................................................... 8 Table 1.2 Parameters in the delivered site-adapted time series and TMY data (hourly time step) .................. 9 Table 2.1 Overview information on the solar meteorological stations installed in Maldives ......................... 10 Table 3.1 Overview information on measurement stations operated in the region ....................................... 12 Table 3.2 Instruments installed at the solar meteorological stations ............................................................ 12 Table 3.3 Technical parameters and accuracy class of the instruments ...................................................... 12 Table 3.4 Overview information on solar meteorological stations operating in the region ............................ 13 Table 3.5 Period of measurements analysed in this report ........................................................................... 13 Table 3.6 Meteorological stations maintenance and instruments field verification ...................................... 13 Table 3.7 Results of field instruments verification at the respective stations ............................................... 14 Table 3.8 Occurrence of data readings for Hanimaadhoo meteorological station ........................................ 14 Table 3.9 Excluded ground measurements after quality control (Sun above horizon) in Hanimaadhoo ....... 15 Table 3.10 Occurrence of data readings for Hulhulé meteorological station .................................................. 16 Table 3.11 Excluded ground measurements after quality control (Sun above horizon) in Hulhulé ................. 16 Table 3.12 Occurrence of data readings for Kadhdhoo meteorological station .............................................. 18 Table 3.13 Excluded ground measurements after quality control (Sun above horizon) in Kadhdhoo ............. 18 Table 3.14 Occurrence of data readings for Gan meteorological station ........................................................ 20 Table 3.15 Excluded ground measurements after quality control (Sun above horizon) in Gan ....................... 20 Table 3.16 Quality control summary (all sites) ................................................................................................ 23 Table 4.1 Input data used in the Solargis and related GHI and DNI outputs for Maldives ............................. 24 Table 4.2 Direct Normal Irradiance: bias and KSI before and after model site-adaptation ............................ 27 Table 4.3 Global Horizontal Irradiance: bias and KSI before and after model site-adaptation ...................... 27 Table 4.4 Direct Normal Irradiance: RMSD before and after model site-adaptation ...................................... 27 Table 4.5 Global Horizontal Irradiance: RMSD before and after model site-adaptation ................................ 27 Table 4.6 Comparison of long term average of yearly summaries of original and site-adapted values ........ 32 Table 5.1 Original source of Solargis meteorological data: models CFSR and CFSv2. .................................. 33 Table 5.2 Solargis meteorological parameters delivered within this project ................................................. 33 Table 5.3 Air temperature at 2 m: accuracy indicators of the model outputs [ºC]. ........................................ 34 Table 5.4 Relative humidity: accuracy indicators of the model outputs [%]. .................................................. 36 Table 5.5 Wind speed: accuracy indicators of the model outputs [m/s]. ....................................................... 39 Table 5.6 Expected uncertainty of modelled meteorological parameters at the project site. ........................ 41 Table 6.1 Uncertainty of the model estimates for original and site-adapted annual long-term values .......... 42 Table 6.2 Annual GHI that should be exceeded with 90% probability in the period of 1 to 10 (25) years ...... 43 Table 6.3 Annual DNI that should be exceeded with 90% probability in the period of 1 to 10 (25) years. ..... 44 Table 6.4 Combined probability of exceedance of annual GHI for uncertainty of the estimate ±3.5%. ......... 45 Table 6.5 Combined probability of exceedance of annual DNI for uncertainty of the estimate ±6.0%. ......... 45 Table 7.1 Parameters in the delivered site-adapted time series and TMY data (hourly time step) ................ 48 Table 7.2 Monthly and yearly long-term GHI averages as calculated from time series and from TMY ......... 49 © 2017 Solargis page 74 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 Table 7.3 Monthly and yearly long-term DNI averages as calculated from time series and from TMY ......... 49 Table 7.4 Monthly and yearly long-term TEMP averages as calculated from time series and from TMY ...... 49 © 2017 Solargis page 75 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 REFERENCES [1] NREL, 1993. User’s Manual for SERI_QC Software-Assessing the Quality of Solar Radiation Data. NREL/TP- 463-5608. Golden, CO: National Renewable Energy Laboratory. [2] Younes S., Claywell R. and Muneer T, 2005. Quality control of solar radiation data: Present status and proposed new approaches. Solar Energy 30, 1533-1549. [3] Perez R., Cebecauer T., Šúri M., 2013. Semi-Empirical Satellite Models. In Kleissl J. (ed.) Solar Energy Forecasting and Resource Assessment. Academic press. http://www.sciencedirect.com/science/article/pii/B9780123971777000024 [4] Cebecauer T., Šúri M., Perez R., High performance MSG satellite model for operational solar energy applications. ASES National Solar Conference, Phoenix, USA, 2010. http://solargis.com/assets/publication/2010/Cebecauer-Suri-Perez--ASES2010--High-performance-MSG- satellite-model-for-operational-solar-energy-applications.pdf [5] Šúri M., Cebecauer T., Perez P., 2010. Quality Procedures of Solargis for Provision Site-Specific Solar Resource Information. Conference SolarPACES 2010, September 2010, Perpignan, France. http://solargis.com/assets/publication/2010//Suri-Cebecaue--Perez-SolarPACES2010--Quality-procedures- of-solargis-for-provision-site-specific-solar-resource-information.pdf [6] Cebecauer T., Suri M., Gueymard C., Uncertainty sources in satellite-derived Direct Normal Irradiance: How can prediction accuracy be improved globally? Proceedings of the SolarPACES Conference, Granada, Spain, 20-23 Sept 2011. [7] http://solargis.com/assets/publication/2011/Cebecauer-Suri-Gueymard--SolarPACES2011--Uncertainty- sources-in-satellite-derived-direct-normal-irradiance--How-can-prediction-accuracy-be-improved-globally.pdf [8] Suri M., Cebecauer T., 2014. Satellite-based solar resource data: Model validation statistics versus user’s uncertainty. ASES SOLAR 2014 Conference, San Francisco, 7-9 July 2014. http://solargis.com/assets/publication/2014//Suri-Cebecauer--ASES-Solar2014--Satellite-Based-Solar- Resource-Data--Model-Validation-Statistics-Versus-User-Uncertainty.pdf [9] Ineichen P., A broadband simplified version of the Solis clear sky model, 2008. Solar Energy, 82, 8, 758-762. [10] Benedictow A. et al. 2012. Validation report of the MACC reanalysis of global atmospheric composition: Period 2003-2010, MACC-II Deliverable D83.1. [11] Molod, A., Takacs, L., Suarez, M., and Bacmeister, J., 2015: Development of the GEOS-5 atmospheric general circulation model: evolution from MERRA to MERRA2, Geosci. Model Dev., 8, 1339-1356, doi:10.5194/gmd- 8-1339-2015 [12] GFS model. http://www.nco.ncep.noaa.gov/pmb/products/gfs/ [13] CFSR model. https://climatedataguide.ucar.edu/climate-data/climate-forecast-system-reanalysis-cfsr/ [14] CFSv2 model http://www.cpc.ncep.noaa.gov/products/CFSv2/CFSv2seasonal.shtml [15] Cano D., Monget J.M., Albuisson M., Guillard H., Regas N., Wald L., 1986. A method for the determination of the global solar radiation from meteorological satellite data. Solar Energy, 37, 1, 31–39. [16] Perez R., Ineichen P., Maxwell E., Seals R. and Zelenka A., 1992. Dynamic global-to-direct irradiance conversion models. ASHRAE Transactions-Research Series, pp. 354-369. [17] Perez, R., Seals R., Ineichen P., Stewart R., Menicucci D., 1987. A new simplified version of the Perez diffuse irradiance model for tilted surfaces. Solar Energy, 39, 221-232. [18] Ruiz-Arias J.A., Cebecauer T., Tovar-Pescador J., Šúri M., 2010. Spatial disaggregation of satellite-derived irradiance using a high-resolution digital elevation model. Solar Energy, 84,1644-57. http://www.sciencedirect.com/science/article/pii/S0038092X10002136 [19] Zelenka A., Perez R., Seals R., Renne D., 1997. Effective Accuracy of Satellite-Derived Hourly Irradiances. Theoretical Appl. Climatology, 62, 199-207. © 2017 Solargis page 76 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 [20] Espinar B., Ramírez L., Drews A., Beyer H. G., Zarzalejo L. F., Polo J., Martín L., Analysis of different comparison parameters applied to solar radiation data from satellite and German radiometric stations. Solar Energy, 83, 1, 118-125, 2009. [21] Lohmann S., Schillings C., Mayer B., Meyer R., 2006. Long-term variability of solar direct and global radiation derived from ISCCP data and comparison with reanalysis data, Solar Energy, 80, 11, 1390-1401. [22] Gueymard C., Solar resource assessment for CSP and CPV. Leonardo Energy webinar, 2010. http://www.leonardo-energy.org/webfm_send/4601 [23] Cebecauer T., Šúri M., 2015. Typical Meteorological Year Data: Solargis Approach. Energy Procedia, Volume 69, 1958-1969. http://dx.doi.org/10.1016/j.egypro.2015.03.195 © 2017 Solargis page 77 of 79 Annual Solar Resource Report for Solar Meteorological Stations in Maldives after completion of 12 months of measurements Solargis reference No. 129-06/2017 SUPPORT INFORMATION Background on Solargis The primary business of the company Solargis is in providing support to the site qualification, planning, financing and operation of solar energy systems. We are committed to increase efficiency and reliability of solar technology by expert consultancy and access to our databases and customer-oriented services. The company builds on experience in solar energy and photovoltaics for 17 years. We strive for development and operation of new generation high-resolution quality-assessed global databases with focus on solar resource and energy-related weather parameters. We are operating simulation, management and control tools, map products, and services for fast access to high quality information needed for system planning, performance assessment, forecasting and management of distributed power generation. ® Company Solargis operates a set of online services, integrated within Solargis information system, which includes data, maps, software, and geoinformation services for solar energy. Legal information Considering the nature of climate fluctuations, interannual and long-term changes, as well as the uncertainty of measurements and calculations, company Solargis cannot take guarantee of the accuracy of estimates. Company Solargis has done maximum possible for the assessment of climate conditions based on the best ® available data, software and knowledge. Solargis is the registered trademark of company Solargis. Other brand names and trademarks that may appear in this study are the ownership of their respective owners. © 2017 Solargis, all rights reserved Solargis is ISO 9001:2008 certified company for quality management. Authors: Marcel Suri Tomas Cebecauer Branislav Schnierer Artur Skoczek Daniel Chrkavy Nada Suriova Maps: Juraj Betak Veronika Madlenakova Project manager: Nada Suriova Approved by: Marcel Suri © 2017 Solargis page 78 of 79