Flooding may damage important transportation infrastructures, such as roads, railways and bridges, which need to be well planned and designed to be able to withstand current and possible future climate-driven increases in flood frequencies and magnitudes. This study develops a novel approach to predictive statistical modelling of the probability of flooding at major road-stream intersection sites, where water, sediment and debris can accumulate and cause failure of drainage facilities and associated road damages. Two areas in south-west Sweden, affected by severe floods in August 2014, are used in representative case studies for this development. A set of physical catchment-descriptors (PCDs), characterizing key aspects of topography, morphology, soil type, land use, hydrology (precipitation and soil moisture) and sediment connectivity in the water-and sediment-contributing catchments, are used for the predictive flood modelling. A main novel contribution to such modelling is to integrate the spatiotemporal characteristics of remotely-sensed soil moisture in indices of sediment connectivity (IC), thereby also allowing for investigation of the role of soil moisture in the flood probability for different road-stream intersections. The results suggest five categories of PCDs as especially important for flood probability quantification and identification of particularly flood-prone intersections along roads (railways, etc.) These include: channel slope at the road-stream intersection and average elevation, soil properties (mainly percentage of till), land use cover (mainly percentage of urban areas), and a sediment connectivity index that considers soil moisture in addition to morphology over the catchment.

Large seasonal variability in precipitation patterns may help overcome data limitations and difficult conditions when characterizing hydrological flow pathways. We used a limited amount of weekly water chemistry as well as stable water isotope data to perform end-member mixing analysis (EMMA) in a generalized likelihood uncertainty estimation (GLUE) framework in a sub-catchment of the Kilombero Valley, Tanzania. While there were considerable uncertainties related to the characterization and mixing of end-members, some robust estimates could be made on contributions to seasonal streamflow variability. For example, there is a low connectivity between the deep groundwater and the stream system throughout the year. Also, a considerable wetting-up period is required before overland flow occurs. Thus, in spite of large uncertainties, our results highlight how improved system understanding of hydrological flows can be obtained even when working in difficult environments.

The expected increase in precipitation and temperature in Scandinavia, and especially short-time heavy precipitation, will increase the frequency of flooding. Urban areas are the most vulnerable, and specifically, the road infrastructure. The accumulation of large volumes of water and sediments on road-stream intersections gets severe consequences for the road drainage structures. This study integrates the spatial and temporal soil moisture properties into the research about flood prediction methods by a case study of two areas in Sweden, Vastra Gotaland and Varmland, which was affected by severe flooding in August 2014. Soil moisture data are derived from remote-sensing techniques, with a focus on the soil moisture-specific satellites ASCAT and SMOS. Furthermore, several physical catchments descriptors (PCDs) are analyzed and the result shows that larger slopes and drainage density, in general, mean a higher risk of flooding. The precipitation is the same; however, it can be concluded that more precipitation in most cases gives higher soil moisture values. The lack, or the dimensioning, of road drainage structures seems to have a large impact on the flood risk as more sediment and water can be accumulated at the road-stream intersection. The results show that the method implementing soil moisture satellite data is promising for improving the reliability of flooding.

This thesis explores key aspects for synthesizing data across spatiotemporal scales relevant for water resources management in an Eastern Africa context. Specifically, the potential of large scale global precipitation datasets (GPDs) in data limited regions to overcome spatial and temporal data gaps is considered. The thesis also explores the potential to utilize limited and non-continuous streamflow and stream water chemistry observations to increase hydrological process understanding. The information gained is then used to build a coherent hydro-climatic framework for streamflow modelling. In this thesis, Kilombero Valley Drainage Basin (KVDB) in Tanzania is used as an example of a data limited region targeted for rapid development, intensification and expansion of agriculture. As such, it is representative for many regions across the Eastern Africa. With regards to the data synthesis, two satellite products, three reanalysis products and three interpolated products were evaluated based on their spatial and temporal precipitation patterns. Streamflow data from KVDB and eight subcatchments were then assessed for quality with regards to missing data. Furthermore, recession analysis was used to estimate catchment-scale characteristic drainage timescale. Results from these streamflow analyses, in conjunction with a hydrological tracer-based analysis, were then used for improved understanding of streamflow generation in the region. Finally, a coherent modelling framework using the HBV rainfall-runoff model was implemented and evaluated based on daily streamflow simulation. Despite the challenges of data limited regions and the often large uncertainty in results, this thesis demonstrates that improved process understanding could be obtained from limited streamflow records and a focused hydrochemical sampling when experimental design natural variability were leveraged to gain a large  signal to noise ratio. Combining results across all investigations rendered information useful for the conceptualization and implementation of the hydro-climatic modelling framework relevant in Kilombero Valley. For example, when synthesized into a coherent framework the GPDs could be downscaled and used for daily streamflow simulations at the catchment scale with moderate success. This is promising when considering the need for estimating impacts of potential future land use and climate change as well as agricultural intensification.

Denna avhandling utforskar aspekter på att syntetisera data med olika rumslig och temporal upplösning, vilket är centralt för vattenförvaltning i östra Afrika. Särskilt fokus ligger på att undersöka möjligheten till att använda globala nederbördsdataset för att fylla rumsliga och temporala luckor där data saknas. Avhandlingen undersökeräven möjligheten till att använda flödesdata med icke-kompletta tidsserier samt kemidata från vattendrag för att utöka kunskap-en om hydrologiska processer. Informationen används för att bygga upp ett integrerande ram-verk för hydro-klimatologisk modellering som exempelvis kan användas för att utforska ef-fekten av ett utökat och intensifierat jordburk på vattenresurser. I denna avhandling användes Kilomberodalens avrinningsområde (Tanzania) som exempel på ett databegränsat område där det pågår en intensiv utökning av jordbruksverksamhet. Detta område kan ses som representa-tivt för ett stort antal områden inom östra Afrika.Datasyntesen innefattade två nederbördsprodukter baserade på satellitdata, tre baserade på återanalysprodukter samt två baserade på interpolering av observervationsdata från regnmä-tare. Dessa åtta produkter utvärderades baserat på deras nederbördsmönster i rum och tid. Ut-över detta utvärderades vattenföringsdata från Kilomberodalens avrinningsområde samt åtta delavrinningsområden utifrån mängden saknad data i respektive tidsserie. Vidare användes resultaten från hydrologisk recessionsanalysför att uppskatta den karaktäristiska avrinningsti-den för avrinningsområden. Resultaten från recessionsanalysensamthydrologiskt spårämnes-försök användessedan för att utöka kunskapen om avrinningsbildning och vattenföring i om-rådet samt som stöd i valet av hydrologiskt modelleringsverktyg. Avslutningsvis användes HBV-avrinningsmodellen för att simulera daglig vattenföring. Trots utmaningen i att arbeta iett databegränsat område och de osäkerheter i resultat som detta tenderar att leda till visar resultaten att det var möjligt att använda begränsad vattenfö-ringsdata och vattenkemidata för att utöka den hydrologiska processförståelsen av området. Detta möjliggjordes genom ett experimentellt upplägg som utnyttjade till ett stort signal-till-brusförhållande under rådande förhållanden av naturlig variabilitet. Kombinerade resultat från alla genomförda studier kunde utnyttjas vid konceptualiseringen och implementeringen av ramverket för hydroklimatologisk modellering av Kilomberodalens avrinningsområde. Till exempel kunde de globala nederbördsdataseten användas för lokal modellering av flödesdata med viss framgång efter syntes och implementering i det integrerande ramverket för hydro-klimatologisk modellering. Detta är lovande med tanke på behovet av att undersöka vilken påverkan möjliga framtida förändringar i markanvändning, klimat samt jordbruk har på den lokala och regionala miljön.

Information on aquifer processes and characteristics across scales has long been a cornerstone for understanding water resources. However, point measurements are often limited in extent and representativeness. Techniques that increase the support scale (footprint) of measurements or leverage existing observations in novel ways can thus be useful. In this study, we used a recession-curve-displacement method to estimate regional-scale aquifer transmissivity (T) from streamflow records across the Kilombero Valley of Tanzania. We compare these estimates to local-scale estimates made from pumping tests across the Kilombero Valley. The median T from the pumping tests was 0.18 m(2)/min. This was quite similar to the median T estimated from the recession-curve-displacement method applied during the wet season for the entire basin (0.14 m(2)/min) and for one of the two sub-basins tested (0.16 m(2)/min). On the basis of our findings, there appears to be reasonable potential to inform water resource management and hydrologic model development through streamflow-derived transmissivity estimates, which is promising for data-limited environments facing rapid development, such as the Kilombero Valley.

This study explored the potential for bias correction of global precipitation datasets (GPD) to support streamflow simulation for water resource management in data limited regions. Two catchments, 580 km(2) and 2530 km(2), in the Kilombero Valley of central Tanzania were considered as case studies to explore three GPD bias correction methods: quantile mapping (QM), daily percentages (DP) and a model based (ModB) bias correction. The GPDs considered included two satellite rainfall products, three reanalysis products and three interpolated observed data products. The rainfall-runoff model HBV was used to simulate streamflow in the two catchments using (1) observed rain gauge data; (2) the original GPDs and (3) the bias-corrected GPDs as input. Results showed that applying QM to bias correction based on limited observed data tends to aggravate streamflow simulations relative to not bias correcting GPDs. This is likely due to a potential lack of representativeness of a single rain gauge observation at the scale of a hydrological catchment for these catchments. The results also indicate that there may be potential benefits in combining streamflow and rain gauge data to bias correct GPDs during the model calibration process within a hydrological modeling framework.

In the face of limited or no precipitation data, global precipitation data sets (GPDs) may provide a viable alternative to gauge or ground radar data. This study aims to provide guidance to the choice of GPDs targeting scales relevant to water resources management in data poor regions. Specifically, the 34 000 km(2) Kilombero Valley in central Tanzania, where water resource management is seen as integral to poverty reduction and food security, is used as a case study for performance evaluation of seven GPDs and their ensemble mean against the Tropical Rainfall Measuring Mission (TRMM) multi-satellite precipitation analysis research-grade product v7 (TRMMv7). The GPDs include one satellite rainfall product [Climate Prediction Center morphing technique v1.0 CRT (CMORPH)], three reanalysis products [Climate Forecasting System Reanalysis (CFSR), European reanalysis interim (ERA-i) and Modern Era Retrospective-Analysis for Research and Applications (MERRA)] and three interpolated data sets [Climate Research Unit Time Series 3.21 (CRU), Global Precipitation and Climatology Center v6 data set (GPCC) and University of Delaware Air Temperature and Precipitation v3.01 data set (UDEL)]. Standard statistical performance measures and spatial patterns were evaluated for the common overlap time period 1998-2010. For this region, the principal seasonality of the climatology was well represented in all GPDs; however, the intraseasonal variability and the spatial precipitation patterns were less well represented. The ensemble mean and GPCC had the best performance with regard to the analysis of the time series while CMORPH and GPCC had the best performance with regard to the spatial pattern analysis. These results indicate that the spatial scale intended for application is a major factor impacting the suitability of a given GPD for hydrometrological studies that form a basis for development of water management strategies.

A history of modern (1973–2015) surface area variability for Lake Basotu in north-central Tanzania has been reconstructed using archived Landsat images from the dry season between June and October. This record was compared to local weather data as well as larger scale weather patterns. The lake has been in a state of decline interrupted by major flood events since the beginning of the satellite record. From 1973 to 1997, the lake area was between 0.97 km² and 4.28 km². Lake extent abruptly increased to 13.86 km² in 1998, when a co-occurrence of El Niño and a positive Indian Ocean Dipole led to extensive flooding. It is hypothesized that local agricultural practices leading to soil erosion and subsequent basin sedimentation have most likely increased the sensitivity of Lake Basotuto climatic fluctuations.

We explore seasonal variability and spatiotemporal patterns in characteristic drainage timescale (K) estimated from river discharge records across the Kilombero Valley in central Tanzania. K values were determined using streamflow recession analysis with a Brutsaert-Nieber solution to the linearized Boussinesq equation. Estimated K values were variable, comparing between wet and dry seasons for the relatively small catchments draining upland positions. For the larger catchments draining through valley bottoms, K values were typically longer and more consistent across seasons. Variations in K were compared with long-term averaged, Moderate-resolution Imaging Spectroradiometer-derived monthly evapotranspiration. Although the variations in K were potentially related to evapotranspiration, the influence of data quality and analysis procedure could not be discounted. As such, even though recession analysis offers a potential approach to explore aquifer release timescales and thereby gain insight to a region's hydrology to inform water resources management, care must be taken when interpreting spatiotemporal shifts in K in connection with process representation in regions like the Kilombero Valley.

Large natural variability in, for example, climate signals and experimental design may help to overcome the data limitations and difficult conditions that typify much of the global south. This, in turn, can facilitate the application of advanced techniques to help inform management with science (which is sorely needed for guiding development). As an example on this concept, we used a limited amount of weekly water chemistry as well as stable water isotope data to perform end-member mixing analysis (EMMA) in a generalized likelihood uncertainty estimation (GLUE) framework in a sub-catchment of Kilombero Valley, Tanzania. How water interacts across the various storages in this region, which has been targeted for rapid agricultural intensification and expansion is still largely unknown, making estimation of potential impacts (not to mention sustainability) associated with various development scenarios difficult. Our results showed that there were, as would be expected, considerable uncertainties related to the characterization of end-members in this remote system. Regardless, some robust estimates could be made on contributions to seasonal streamflow variability. For example, it appears that there is a low connectivity between the deep groundwater and the stream system throughout the year. Also, there is a considerable wetting up period required before overland flow occurs. These process insights, in turn, help interpreting hydrochemical data thereby potentially improving understanding at larger scales. Thus, in spite of large uncertainties our results highlight how improved system understanding of hydrological flows can be obtained even when working under less than perfect conditions.