Mountains in the Tibetan Plateau known as Asian water towers benefit more than one sixth of the population of the world. Runoff from high mountains is indispensable for domestic water supply, irrigation, hydropower production and industrial purposes. Although the importance of mountainous areas has been touched upon by some benchmark reports, the hydrological processes are still poorly characterized. Distributed physically-based hydrological models are essential to better understand hydrological processes, and numerous models of varying complexity have been used in mountainous areas for forecasting runoff, snow-cover evolution, and related processes. Precipitation is one of the most critical model inputs for accurate hydrological simulation. However, characterization of precipitation in high altitude localities is a challenging task because the hydrometeorological data sets are extremely limited and precipitation is highly variable in mountainous areas.
Ph.D Candidates Li Wang, a member of Professor Fan Zhang’s team of the CAS center for excellence in Tibetan Plateau earth sciences and the Institute of Tibetan Plateau Research, and the co-authors, find out that precipitation is greatly influenced by altitude in the study area based on intensive observation of precipitation in the Mabengnong catchment in the southeast of the Tibetan Plateau during July to August 2013. The average Precipitation Gradient (PG) is 0.10, 0.28 and 0.26 mm/d/100m (Figure 1) and the average duration is around 0.1, 0.8 and 5.2 hours for trace, light and moderate rain, respectively. A distributed biosphere hydrological model based on water and energy budgets with improved physical process for snow (WEB-DHM-S) is applied to simulate the hydrological processes with gridded precipitation data derived from a lower altitude meteorological station and the PG and hourly distribution (HD) characterized for the study area. The observed runoff, MODIS/Terra snow cover area (SCA) data, and MODIS/Terra land surface temperature (LST) data are used for model calibration and validation. Runoff, SCA and LST simulations all show reasonable results (Figure 2). Sensitivity analyses illustrate that runoff is largely underestimated without considering PG, indicating that short-term intensive precipitation observation has the potential to greatly improve hydrological modelling of poorly gauged high mountain catchments.

Figure 1 Precipitation gradients of (a) trace, (b) light and (c) moderate precipitation

Figure 2 Comparision of observed and simulated runoff, SCA and LST. Base case used the PG and HD derived based on observation; Case 1 used the same PG as the base case but uniform distribution of precipitation within the day; Case 2 used PG of zero but the same HD as the base case; Case 3 used PG of zero and uniform distribution of precipitation within the day.

This paper has been published in Journal of Hydrology with paper link: https://doi.org/10.1016/j.jhydrol.2017.11.039 . Citation: Li Wang, Fan Zhang, Hongbo Zhang, Christopher A. Scott, Chen Zeng, Xiaonan Shi. Intensive precipitation observation greatly improves hydrological modelling of the poorly gauged high mountain Mabengnong catchment in the Tibetan Plateau. Journal of Hydrology,2018,556:500-509.