Hydrological Modelling of Mountainous and Glacierised regions under Changing Climate

Abstract

Climate change is one of the most serious environmental threats that humanity has ever been confronted to. Hydrological models are vital tools to asses its impacts on the water cycle and water resources. The goal of this project is to evaluate and improve the capacity of the HBV model (Hydrologiska Byr°ans Vattenbalansavdelning) in simulating hydrological processes in mountainous and glacierised regions under both the present and future climate. This goal is achieved in two steps: (1) implement routing and glacier retreating modules in a grid-based HBV model for mountainous and glacierised areas in Norway and the Himalayan region (i.e. Bhutan and India); (2) calculate available water resource scenarios for two Himalayan basins with considerable glacier coverage and growing water demand. The thesis is composed of four peer-reviewed articles which focus on the two steps. Articles I and II examine model setting effects on their performance in reproducing major hydrological processes. Articles I implemented routing algorithms in the grid-based HBV model and tested them at various spatial resolutions in the Glomma basin (Norway). The routing procedures improved the model performance in daily runoff simulation to varying extents. A hillslope routing method and its combination with the channel routing returned a Nash-Sutcliffe coefficient 0.05 higher than the initial grid-based model. Articles II evaluated five variants of the HBV model in runoff simulation of interior points and internal variables in the Norsfoss basin (Norway). The five variants were a lumped (LWhole), a semi-distributed (SBand) and three grid-based models, GRZero (without routing), GROne (hillslope routing) and GRTwo (hillslope and channel routing). For runoff simulations, GRTwo and GROne were superior over other model variants in model efficiency, particularly in simulating low flow. This superiority deteriorated in reproducing runoff at six interior points. Compared with measurements of snow water equivalent at snow pillows and groundwater depth in piezometers, all grid-based models had a similar efficiency. Articles IIII and IV include an integration of a mass conserving glacier model into the HBV model and its application in water resource projections. Article III combined the glacier retreat model with the HBV model. The coupled model was tested in three basins (the Nigardsbreen basin in Norway, the Chamkhar Chhu basin in Bhutan and the Beas basin in India) with different glacier coverage and hydrologic regime. Results showed that, in addition to runoff simulation, the model gave a high efficiency in reproducing glacier annual mass balance in the Nigardsbreen basin where measurements are available to verify the results. Moreover, the model provided maps of snow distribution and glacier runoff. Article IV projected available water resources per capita (Wp) for the Chamkhar Chhu (eastern Himalaya) and the Beas (western Himalaya) basins for the period 2010–2050. All climate projections indicated significant increases in annual temperature, but not in annual precipitation. All Wp projections revealed pronounced water resources drops jointly induced by continuous climate change and population growth. The latter is responsible for roughly 40% of the water declines. The regional climate models and CO2 emissions cause approximately 30% unceratainties ranging from 11% to 44%. When considering ±20% inaccuracy in population estimations, the highest uncertainty reaches 87%. The uncertainties are worthy of attention, but there is no doubt that the two basins are facing serious water scarcity and the water conditions will get worse. Water shortage has been and continues to be a major constraint of economic and social development.