ESTIMATION OF SNOW MELT WATER USING REMOTE SENSING AND SNOW PROCESS MODEL OVER THE QINGHAI-TIBET PLATEAU

J. Shi^1,2, J. Pan³, C. Xiong⁴, L. Jiang⁵

¹Chinese Academy of Sciences, 2National Space Science Center, ³National Space Science Center CAS, ⁴Southwest Jiaoton University, ⁵Beijing Normal University

Snowmelt plays a critical role in shaping hydrological processes in cold regions. During the snow-covered period, the snowpack serves as both the upper infiltration and thermal boundaries for frozen soil. As the “Water Tower of Asia,” the Qinghai-Tibet Plateau (QTP) is endowed with abundant mid-latitude snow and ice meltwater and frozen soil resources. However, the station measurements are sparse and non-representative in this region. Under the QTP’s complex terrain and intense solar radiation, snowpacks exhibit deep and heterogeneous distributions in mountainous areas, while flat regions are characterized by discontinuous and rapidly changing snow cover. Therefore, Estimating key snow cover variables—including areal extent, snow depth, snow water equivalent (SWE), snowmelt phenology, and snowmelt runoff—remains highly uncertain in both observational and modeling frameworks.

Remote sensing techniques are essential for addressing the data scarcity issues. We conducted a systematic research on the QTP snowpack by integrating remote sensing and snow process modeling, with the goal of analyzing the spatiotemporal distribution and variation characteristics of seasonal snow cover, snow water equivalent and snowmelt.

On Remote sensing aspect, we have developed the algorithms 1) to estimate daily fraction snow cover from optical sensors of polar-orbiting and geostationary satellites. This method effectively reduces the influence of cloud cover and mixed pixels, enabling sub-pixel fraction snow cover inversion at a daily scale; 2) to estimate snow depth using microwave radiometer sensors. We will demonstrate the algorithms and validations in details. 3) It was found that the significant difference between the maximum SWE and amount of snow melting water. This is because the discontinuous snowpack on the time scale and new snow fall during melting season, where microwave observations failed to estimate snow depth. Therefore, we designed a nested-grid observation framework to integrate remote sensing data across two scales. By physically coupling the SNTHERM snow process model with the microwave snow emission model, we solved the observation equation based on adjustment computation theory, for snow depth, snow water equivalent (SWE), and snowmelt dynamics. This framework enables: (1) constraining snow mass in coarse-resolution grids using passive microwave observations; and; (2) leveraging optical remote sensing to enhance spatial resolution and resolve snowmelt processes during periods when passive microwave data are unreliable.

We generated a 20-year, 3-km dataset of multiple snow parameters over the Qinghai-Tibet Plateau (QTP), which has been available in the National Tibetan Plateau Data Center (NTPDC) repository (https://data.tpdc.ac.cn/zh-hans/data/4ebaef2a-a0c1-48d7-a19c-ebf19fea2167). By incorporating water balance constraints into the energy balance framework, we clarified the contributions of various components of the snow water balance to snowmelt runoff and analyzed the spatiotemporal distribution, trends, and regional heterogeneity of snow cover parameters across the plateau. These systematic advancements in observational capacity have yielded the most reliable and parameter complete snow datasets currently available, providing robust support for scientific research on the impacts of QTP snow under climatic change impact in this region.