Projections for headwater catchments of the Tarim River reveal glacier retreat and decreasing surface water availability but uncertainties are large [article]

Doris Duethmann, Christoph Menz, Tong Jiang, Sergiy Vorogushyn, University, My, University, My
In the Tarim River Basin, water resources from the mountain areas play a key role due to the extremely arid climate of the lowlands. This study presents an analysis of future climate change impacts on glaciers and surface water availability for headwater catchments of the Aksu River, the most important tributary to the Tarim River. We applied a glacio-hydrological model that underwent a comprehensive multivariable and multiobjective model calibration and evaluation, based on daily and
more » ... l discharge variations and glacier mass changes. Transient glacier geometry changes are simulated using the Δh-approach. For the ensemble-based projections, we considered three different emission scenarios, nine global climate models (GCMs) and two regional climate models, and different hydrological model parameters derived from the multiobjective calibration. The results show a decline in glacier area of −90% to −32% until 2099 (reference ~2008) (based on the 5–95 percentile range of the ensemble). Glacier melt is anticipated to further increase or stay at a high level during the first decades of the 21st century, but then declines because of decreased glacier extents. Overall discharge in the Aksu headwaters is expected to be increased in the period 2010–2039 (reference 1971–2000), but decreased in 2070–2099. Seasonally, projections show an increase in discharge in spring and early summer throughout the 21st century. Discharge changes in mid to late summer are more variable, with increases or decreases depending on the considered period and GCM. Uncertainties are largely caused by differences between the different GCMs, with further important contributions from different emission scenarios in the second half of the 21st century. Contributions from the hydrological model parameters to the ensemble uncertainty were generally found to be small.
doi:10.34657/347 fatcat:4hnztefhmjggff4rv7qnbddusq