Master of Arts / Department of Geography / Jida Wang / In the 1950s-60s, the world was experiencing a dramatic increase in artificial water impoundments in an unprecedented effort to eliminate spatial and temporal variations in water accessibility. Dam construction has since decreased, but recently, efforts to supply water and power to a booming world population may once again spark another dam construction boom. Water supplies in some regions are already highly stressed in an effort to satisfy the ever-growing water demand for agriculture, industrial, and domestic uses. With nearly 30% of the world’s population living in water-scarce regions, an improved understanding of total stored surface water has never been more needed. Yet, a complete and spatially-explicit, worldwide inventory of such storage capabilities is lacking. Using several open-source dam registries and high-resolution global lake mapping datasets extracted from thousands of Landsat images, we here aim to provide an updated and spatially-explicit inventory of dams and artificial reservoirs across the world. The following research uses novel techniques to merge 5 authoritative, open-source dam registries into a single dam and reservoir dataset, which we deemed as the Global Dam and Reservoir Inventory (GDRI). In total, GDRI documents 89,500 dams and 83,767 reservoirs for a total capacity of 8,492km3 and total surface area of 754,551km2. Reservoirs account for approximately 2.5% of the Earth’s terrestrial water. In other words, 1 unit of water for every 40 units has been artificially created. Further downscaling of the non-geocoded records provided by the International Commission of Large Dams (ICOLD) using similar geocoding methods allowed for the thorough use of all available ICOLD records. Additional capacity estimates from downscaled ICOLD records increased the GDRI capacity documentation to 8,603km3 and surface area documentation to 859,271km2. Compared to its counterpart, the Global Reservoir and Dam dataset (GRanD), GDRI increased the number of dams documented by 1204%, reservoirs by 1127%, total capacity by 37%, and total surface area by 68%. Initial water impoundment from dam construction activities can lower sea level by permanently trapping water storage on land. Dam construction resulted in an equivalent sea level drop (SLD) of 23.4mm or 0.08mm/yr. Since the dam construction boom of the 1950s-1960s, yearly SLD increased to 0.27mm/yr. By considering the hydrological characteristics of dam location, in terms of endorheic and exorheic basins, we found that exclusion of endorheic located dams decreases the overall effect on SLD by 5.47% or 1.28mm. Failure to consider the hydrologic characteristic of dam location can result in the overestimation of dam-induced SLD. After the dam construction boom of the 1950s-1960s, the world has seen a decreasing trend in dam construction, but developing countries (China, Brazil, India) are still actively pursuing dam projects that are larger and more ambitious than ever before. We see less developed countries often lack the capabilities for dam construction possibly increasing stress on natural water supplies in those regions. The datasets produced are by no means perfect. Overall, the described procedures should be considered a heuristic model, where fastidious quality assurance and automated procedures work to thoroughly eliminate many of the issues encountered with the dataset production, but errors may still exist. However, duplication between the contributing dam datasets, spatial limitations of the lake datasets, imperfect geocoding procedures, and inclusion of more dam datasets provide opportunity for future refinement and improvement of the datasets. This research contributes vital information about anthropogenic water resources that incrementally enhances our knowledge of global hydrology and the interactions taking place between different water entities.

Water is essential for human well-being and healthy ecosystems, but population growth and changes in climate and land-use are putting increased stress on water resources in many regions. To ensure water security, knowledge about the spatiotemporal distribution of these resources is of great importance. However, estimates of global water resources are constrained by limitations in availability and quality of data. This thesis explores the quality of both observational and modelled data, gives an overview of models used for large-scale hydrological modelling, and explores the possibilities to deal with the scarcity of data by prediction of flow-duration curves. The evaluation of the quality of observational data for large-scale hydrological modelling was based on both hydrographic data, and model forcing and evaluation data for basins worldwide. The results showed that a GIS polygon dataset outperformed all gridded hydrographic products analysed in terms of representation of basin areas. Through a screening methodology based on the long-term water-balance equation it was shown that as many as 8–43% of the basins analysed displayed inconsistencies between forcing (precipitation and potential evaporation) and evaluation (discharge) data depending on how datasets were combined. These data could prove disinformative in hydrological model inference and analysis. The quality of key hydrological variables from a numerical weather prediction model was assessed by benchmarking against observational datasets and by analysis of the internal land-surface water budgets of several different model setups. Long-term imbalances were found between precipitation and evaporation on the global scale and between precipitation, evaporation and runoff on both cell and basin scales. These imbalances were mainly attributed to the data assimilation system in which soil moisture is used as a nudge factor to improve weather forecasts. Regionalisation, i.e. transfer of information from data-rich areas to data-sparse areas, is a necessity in hydrology because of a lack of observed data in many areas. In this thesis, the possibility to predict flow-duration curves in ungauged basins was explored by testing several different methodologies including machine learning. The results were mixed, with some well predicted curves, but many predicted curves exhibited large biases and several methods resulted in unrealistic curves. / Vatten är en förutsättning för människors och ekosystems hälsa, men befolkningsökning och förändringar av klimat och markanvändning förväntas öka trycket på vattenresurserna i många regioner i världen. För att kunna säkerställa en god tillgång till vatten krävs kunskap om hur dessa resurser varierar i tid och rum. Tillförlitligheten hos skattningar av globala vattenresurser begränsas dock både av begränsad tillgänglighet av och kvalitet hos observerade data. Denna avhandling utforskar kvaliteten av såväl observations- som modellbaserade data, ger en överblick över modeller som används för storskalig hydrologisk modellering och utforskar möjligheterna att förutsäga varaktighetskurvor som ett sätt att hantera bristen på data i många områden. Utvärderingen av observationsbaserade datas kvalitet baserades på hydrografiska data och driv- och utvärderingsdata för storskaliga hydrologiska modeller. Resultaten visade att en uppsättning data över hydrografin baserad på GIS-polygoner representerade avrinningsområdesareorna bättre än alla de som byggde på rutor. En metod baserad på långtidsvattenbalansen identifierade att kombinationen av drivdata (nederbörd och potentiell avdunstning) och utvärderingsdata (vattenföring) var fysiskt orimlig för så många som 8–43 % av de analyserade avrinningsområdena beroende på hur olika datauppsättningar kombinerades. Sådana data kan vara desinformativa för slutsatser som dras av resultat från hydrologiska modeller och analyser. Kvaliteten hos hydrologiskt viktiga variabler från en numerisk väderprognosmodell utvärderades dels genom jämförelser med observationsdata och dels genom analys av landytans vattenbudget för ett flertal olika modellvarianter. Resultaten visade obalanser mellan långtidsvärden av nederbörd och avdunstning i global skala och mellan långtidsvärden av nederbörd, avdunstning och avrinning i både modellrute- och avrinningsområdesskala. Dessa obalanser skulle till stor del kunna förklaras av den data assimilering som görs, i vilken markvattenlagret används som en justeringsfaktor för att förbättra väderprognoserna. Regionalisering, som innebär en överföring av information från områden med god tillgång på mätdata till områden med otillräcklig tillgång, är i många fall nödvändig för hydrologisk analys på grund av att mätdata saknas i många områden. I denna avhandling utforskades möjligheten att förutsäga varaktighetskurvor för avrinningsområden utan vattenföringsdata genom flera metoder inklusive maskininlärning. Resultaten var blandade med en del kurvor som förutsas väl, och andra kurvor som visade stora systematiska avvikelser. Flera metoder resulterade i orealistiska kurvor (ickemonotona eller med negativa värden).