{"id":{"repo_id":"stellenbosch","oai_identifier":"oai:scholar.sun.ac.za:10019.1/135974"},"canonical_url":"https://search.dev.ndltd.org/etd/stellenbosch/oai:scholar.sun.ac.za:10019.1/135974","repository":{"repo_id":"stellenbosch","name":"Stellenbosch University","base_url":"https://scholar.sun.ac.za/server/oai/request"},"display":{"title":"The development of Intensity Duration Frequency curves under climate change considerations using Remote Sensed rainfall data in Uganda","abstract":"The intensification of extreme rainfall poses significant risks to hydrological infrastructure, particularly in data-scarce regions such as Uganda. Reliable intensity–duration–frequency (IDF) curves are essential for climate-resilient design, yet their development is constrained by the lack of long-term subdaily rainfall records. The challenge addressed in this research is the limited availability of subdaily and long-term daily rainfall data required to generate IDF curves. The research addressed this challenge by evaluating and bias-correcting remotely sensed rainfall (RSR) data, assessing daily rainfall disaggregation methods, and integrating observed rainfall, bias-corrected RSR, and regional climate model (RCM) outputs to develop IDF curves for both the current (1991–2020) and mid-century (2036–2065) climate scenarios. Seven RSR datasets were evaluated against observed records using the root mean square error (RMSE), percent bias (PBIAS), and Kolmogorov–Smirnov metrics. The National Oceanic and Atmospheric Administration–Climate Prediction Centre (NOAA_CPC) product performed best at Jinja (PBIAS = –12.95%), whereas the Global Precipitation Climatology Centre (GPCC) dataset showed better performance at Soroti (–9.66%), Mbarara (–5.93%), and Gulu (–3.05%). To further improve their accuracy, the datasets were bias-corrected using four different methods, among which quantile mapping (QM) emerged as the most effective. For example, at Gulu, QM reduced the NOAA_CPC RMSE from 29.20 mm to 19.00 mm and improved the PBIAS from –19.23% to 1.05%. IDF curves developed from observed, bias-corrected RSR, and best-evaluated RCMs (BCCRWRF331 and REMO2009) were fitted with the generalized extreme value distribution. The results revealed strong spatial variability from RCM outputs. Under representative concentration pathway (RCP) 4.5, intensities increased moderately at Gulu (9–21%) but rose at Jinja (50–79%) and Mbarara (46–73%), with Fort Portal showing the greatest increase (up to 170% for rare events). Soroti showed increases at low return periods but decreases at higher return periods. Under RCP8.5, most stations recorded increased intensities, whereas Jinja and Gulu showed decreases, reflecting localized climate variability. To support application in ungauged catchments, multilinear regression analysis (MLR) models were then developed to regionalize IDF coefficients based on elevation and the mean of annual maximum series (AMS) rainfall as predictors. Station-level validation using leave-oneout cross-validation (LOOCV) revealed high numerical accuracy, with percentage errors below 2% at Tororo, Arua, and Soroti for coefficient b and below 5% for all stations for c. In conclusion, bias-corrected RSR data and the aligned time distribution (TD) disaggregation approach offer reliable alternatives for IDF curve development in data-scarce settings. Novel contributions include a bias correction framework and a MLR regionalization model. These tools support climate-resilient hydraulic infrastructure planning in data-scarce regions. It is recommended that government agencies, along with planners and consultants, consider using bias-corrected RSR datasets in areas lacking ground-based rainfall and use aligned TD factors where subdaily records are unavailable. Infrastructure design should integrate IDF curves under both RCP4.5 and RCP8.5 scenarios, with design standards updated to support scenario-based planning that accounts for a range of plausible climate futures.","abstract_html":"The intensification of extreme rainfall poses significant risks to hydrological infrastructure, particularly in data-scarce regions such as Uganda. Reliable intensity–duration–frequency (IDF) curves are essential for climate-resilient design, yet their development is constrained by the lack of long-term subdaily rainfall records. The challenge addressed in this research is the limited availability of subdaily and long-term daily rainfall data required to generate IDF curves. The research addressed this challenge by evaluating and bias-correcting remotely sensed rainfall (RSR) data, assessing daily rainfall disaggregation methods, and integrating observed rainfall, bias-corrected RSR, and regional climate model (RCM) outputs to develop IDF curves for both the current (1991–2020) and mid-century (2036–2065) climate scenarios. Seven RSR datasets were evaluated against observed records using the root mean square error (RMSE), percent bias (PBIAS), and Kolmogorov–Smirnov metrics. The National Oceanic and Atmospheric Administration–Climate Prediction Centre (NOAA_CPC) product performed best at Jinja (PBIAS = –12.95%), whereas the Global Precipitation Climatology Centre (GPCC) dataset showed better performance at Soroti (–9.66%), Mbarara (–5.93%), and Gulu (–3.05%). To further improve their accuracy, the datasets were bias-corrected using four different methods, among which quantile mapping (QM) emerged as the most effective. For example, at Gulu, QM reduced the NOAA_CPC RMSE from 29.20 mm to 19.00 mm and improved the PBIAS from –19.23% to 1.05%. IDF curves developed from observed, bias-corrected RSR, and best-evaluated RCMs (BCCRWRF331 and REMO2009) were fitted with the generalized extreme value distribution. The results revealed strong spatial variability from RCM outputs. Under representative concentration pathway (RCP) 4.5, intensities increased moderately at Gulu (9–21%) but rose at Jinja (50–79%) and Mbarara (46–73%), with Fort Portal showing the greatest increase (up to 170% for rare events). Soroti showed increases at low return periods but decreases at higher return periods. Under RCP8.5, most stations recorded increased intensities, whereas Jinja and Gulu showed decreases, reflecting localized climate variability. To support application in ungauged catchments, multilinear regression analysis (MLR) models were then developed to regionalize IDF coefficients based on elevation and the mean of annual maximum series (AMS) rainfall as predictors. Station-level validation using leave-oneout cross-validation (LOOCV) revealed high numerical accuracy, with percentage errors below 2% at Tororo, Arua, and Soroti for coefficient b and below 5% for all stations for c. In conclusion, bias-corrected RSR data and the aligned time distribution (TD) disaggregation approach offer reliable alternatives for IDF curve development in data-scarce settings. Novel contributions include a bias correction framework and a MLR regionalization model. These tools support climate-resilient hydraulic infrastructure planning in data-scarce regions. It is recommended that government agencies, along with planners and consultants, consider using bias-corrected RSR datasets in areas lacking ground-based rainfall and use aligned TD factors where subdaily records are unavailable. Infrastructure design should integrate IDF curves under both RCP4.5 and RCP8.5 scenarios, with design standards updated to support scenario-based planning that accounts for a range of plausible climate futures.","abstract_has_math":false,"creators":["Okirya, Martin"],"institution":"Stellenbosch : Stellenbosch University","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Du Plessis, Jakobus A."],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-03","date_published":"2026-03","updated_at":"2026-07-24T04:40:09Z","subjects":[],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://scholar.sun.ac.za/handle/10019.1/135974","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Du Plessis, Jakobus A."]},{"key":"dc:contributor.other","label":"Dc Contributor Other","values":["Stellenbosch University. Faculty of Engineering. Dept. of Civil Engineering."]},{"key":"dc:creator","label":"Author","values":["Okirya, Martin"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-04-16T14:06:18Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-04-16T14:06:18Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-03"]},{"key":"dc:publisher","label":"Institution","values":["Stellenbosch : Stellenbosch University"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://scholar.sun.ac.za/handle/10019.1/135974"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (PhD)--Stellenbosch University, 2026.","Okirya, M. 2026. The development of Intensity Duration Frequency curves under climate change considerations using Remote Sensed rainfall data in Uganda. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/761fbadd-675c-4e23-a2f3-a6b342a08ea0"]},{"key":"dc:description.abstract","label":"Abstract","values":["The intensification of extreme rainfall poses significant risks to hydrological infrastructure, particularly in data-scarce regions such as Uganda. Reliable intensity–duration–frequency (IDF) curves are essential for climate-resilient design, yet their development is constrained by the lack of long-term subdaily rainfall records. The challenge addressed in this research is the limited availability of subdaily and long-term daily rainfall data required to generate IDF curves. The research addressed this challenge by evaluating and bias-correcting remotely sensed rainfall (RSR) data, assessing daily rainfall disaggregation methods, and integrating observed rainfall, bias-corrected RSR, and regional climate model (RCM) outputs to develop IDF curves for both the current (1991–2020) and mid-century (2036–2065) climate scenarios. Seven RSR datasets were evaluated against observed records using the root mean square error (RMSE), percent bias (PBIAS), and Kolmogorov–Smirnov metrics. The National Oceanic and Atmospheric Administration–Climate Prediction Centre (NOAA_CPC) product performed best at Jinja (PBIAS = –12.95%), whereas the Global Precipitation Climatology Centre (GPCC) dataset showed better performance at Soroti (–9.66%), Mbarara (–5.93%), and Gulu (–3.05%). To further improve their accuracy, the datasets were bias-corrected using four different methods, among which quantile mapping (QM) emerged as the most effective. For example, at Gulu, QM reduced the NOAA_CPC RMSE from 29.20 mm to 19.00 mm and improved the PBIAS from –19.23% to 1.05%. IDF curves developed from observed, bias-corrected RSR, and best-evaluated RCMs (BCCRWRF331 and REMO2009) were fitted with the generalized extreme value distribution. The results revealed strong spatial variability from RCM outputs. Under representative concentration pathway (RCP) 4.5, intensities increased moderately at Gulu (9–21%) but rose at Jinja (50–79%) and Mbarara (46–73%), with Fort Portal showing the greatest increase (up to 170% for rare events). Soroti showed increases at low return periods but decreases at higher return periods. Under RCP8.5, most stations recorded increased intensities, whereas Jinja and Gulu showed decreases, reflecting localized climate variability. To support application in ungauged catchments, multilinear regression analysis (MLR) models were then developed to regionalize IDF coefficients based on elevation and the mean of annual maximum series (AMS) rainfall as predictors. Station-level validation using leave-oneout cross-validation (LOOCV) revealed high numerical accuracy, with percentage errors below 2% at Tororo, Arua, and Soroti for coefficient b and below 5% for all stations for c. In conclusion, bias-corrected RSR data and the aligned time distribution (TD) disaggregation approach offer reliable alternatives for IDF curve development in data-scarce settings. Novel contributions include a bias correction framework and a MLR regionalization model. These tools support climate-resilient hydraulic infrastructure planning in data-scarce regions. It is recommended that government agencies, along with planners and consultants, consider using bias-corrected RSR datasets in areas lacking ground-based rainfall and use aligned TD factors where subdaily records are unavailable. Infrastructure design should integrate IDF curves under both RCP4.5 and RCP8.5 scenarios, with design standards updated to support scenario-based planning that accounts for a range of plausible climate futures."]},{"key":"dc:title","label":"Title","values":["The development of Intensity Duration Frequency curves under climate change considerations using Remote Sensed rainfall data in Uganda"]}]}],"canonical_facts":{"dc:contributor.advisor":["Du Plessis, Jakobus A."],"dc:contributor.other":["Stellenbosch University. Faculty of Engineering. Dept. of Civil Engineering."],"dc:creator":["Okirya, Martin"],"dc:date.accessioned":["2026-04-16T14:06:18Z"],"dc:date.available":["2026-04-16T14:06:18Z"],"dc:date.issued":["2026-03"],"dc:description":["Thesis (PhD)--Stellenbosch University, 2026.","Okirya, M. 2026. The development of Intensity Duration Frequency curves under climate change considerations using Remote Sensed rainfall data in Uganda. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/761fbadd-675c-4e23-a2f3-a6b342a08ea0"],"dc:description.abstract":["The intensification of extreme rainfall poses significant risks to hydrological infrastructure, particularly in data-scarce regions such as Uganda. Reliable intensity–duration–frequency (IDF) curves are essential for climate-resilient design, yet their development is constrained by the lack of long-term subdaily rainfall records. The challenge addressed in this research is the limited availability of subdaily and long-term daily rainfall data required to generate IDF curves. The research addressed this challenge by evaluating and bias-correcting remotely sensed rainfall (RSR) data, assessing daily rainfall disaggregation methods, and integrating observed rainfall, bias-corrected RSR, and regional climate model (RCM) outputs to develop IDF curves for both the current (1991–2020) and mid-century (2036–2065) climate scenarios. Seven RSR datasets were evaluated against observed records using the root mean square error (RMSE), percent bias (PBIAS), and Kolmogorov–Smirnov metrics. The National Oceanic and Atmospheric Administration–Climate Prediction Centre (NOAA_CPC) product performed best at Jinja (PBIAS = –12.95%), whereas the Global Precipitation Climatology Centre (GPCC) dataset showed better performance at Soroti (–9.66%), Mbarara (–5.93%), and Gulu (–3.05%). To further improve their accuracy, the datasets were bias-corrected using four different methods, among which quantile mapping (QM) emerged as the most effective. For example, at Gulu, QM reduced the NOAA_CPC RMSE from 29.20 mm to 19.00 mm and improved the PBIAS from –19.23% to 1.05%. IDF curves developed from observed, bias-corrected RSR, and best-evaluated RCMs (BCCRWRF331 and REMO2009) were fitted with the generalized extreme value distribution. The results revealed strong spatial variability from RCM outputs. Under representative concentration pathway (RCP) 4.5, intensities increased moderately at Gulu (9–21%) but rose at Jinja (50–79%) and Mbarara (46–73%), with Fort Portal showing the greatest increase (up to 170% for rare events). Soroti showed increases at low return periods but decreases at higher return periods. Under RCP8.5, most stations recorded increased intensities, whereas Jinja and Gulu showed decreases, reflecting localized climate variability. To support application in ungauged catchments, multilinear regression analysis (MLR) models were then developed to regionalize IDF coefficients based on elevation and the mean of annual maximum series (AMS) rainfall as predictors. Station-level validation using leave-oneout cross-validation (LOOCV) revealed high numerical accuracy, with percentage errors below 2% at Tororo, Arua, and Soroti for coefficient b and below 5% for all stations for c. In conclusion, bias-corrected RSR data and the aligned time distribution (TD) disaggregation approach offer reliable alternatives for IDF curve development in data-scarce settings. Novel contributions include a bias correction framework and a MLR regionalization model. These tools support climate-resilient hydraulic infrastructure planning in data-scarce regions. It is recommended that government agencies, along with planners and consultants, consider using bias-corrected RSR datasets in areas lacking ground-based rainfall and use aligned TD factors where subdaily records are unavailable. Infrastructure design should integrate IDF curves under both RCP4.5 and RCP8.5 scenarios, with design standards updated to support scenario-based planning that accounts for a range of plausible climate futures."],"dc:identifier.uri":["https://scholar.sun.ac.za/handle/10019.1/135974"],"dc:language.iso":["en"],"dc:publisher":["Stellenbosch : Stellenbosch University"],"dc:title":["The development of Intensity Duration Frequency curves under climate change considerations using Remote Sensed rainfall data in Uganda"],"dc:type":["Thesis"]},"updated_at":"2026-07-24T04:40:09Z"}