Snow Cover Phenology in Xinjiang Based on a Novel Method and MOD10A1 Data

Abstract

Using Earth observation to accurately extract snow phenology changes is of great significance for deepening the understanding of the ecological environment and hydrological process, agricultural and animal husbandry production, and high-quality development of the social economy in Xinjiang. Considering snow cover phenology based on MODIS product MOD10A1 data, this paper constructed a method for automatically extracting key phenological parameters in Xinjiang and calculated three key phenological parameters in Xinjiang from 2001 to 2020, including SCD (snow cover duration), SOD (snow onset date), and SED (snow end date). The daily data of four field camera observation points during an overlapping period from 2017 to 2019 were used to evaluate the snow cover phenological parameters extracted by MOD10A1, and the mean absolute error (MAE) and root mean square error (RMSE) values were 0.65 and 1.07, respectively. The results showed the following: 1. The spatiotemporal variation in snow phenology was highly altitude dependent. The mean gradients of SCD in the Altai Mountains, Tienshan Mountains, and Kunlun Mountains is 2.6, 2.1, and 1.2 d/100 m, respectively. The variation trend of snow phenology with latitude and longitude was mainly related to the topography of Xinjiang. Snow phenological parameters of different land-use types were different. The SCD values in wasteland were the lowest and the SED was the earliest, while forest land was the first to enter SOD accumulation. According to the study, the mean annual values of SCD, SOD, and SED were 25, 342 (8 December), and 51 (8 February) as day of year (DOY), respectively. The snow cover area was mainly distributed in the Altai Mountains, Junggar Basin, Tianshan Mountains, and Kunlun Mountains. 2. The variation trend and significance of snow cover phenological parameters in different regions are different, and the variation trend of snow cover phenological parameters in most regions of Xinjiang is non-significant.