Vegetation expansion in the subnival Hindu Kush Himalaya

The mountain systems of the Hindu Kush Himalaya (HKH) are changing rapidly due to climatic change, but an overlooked component is the subnival ecosystem (between the treeline and snow line), characterized by short-stature plants and seasonal snow. Basic information about subnival vegetation distribution and rates of ecosystem change are not known, yet such information is needed to understand relationships between subnival ecology and water/carbon cycles. We show that HKH subnival ecosystems cover five to 15 times the area of permanent glaciers and snow, highlighting their eco-hydrological importance. Using satellite data from the Landsat 5, 7 and 8 missions, we measured change in the spatial extent of subnival vegetation from 1993 to 2018. The Landsat surface reflectance-derived Normalized Difference Vegetation Index product was thresholded at 0.1 to indicate the presence/absence of vegetation. Using this product, the strength and direction of time-series trends in the green pixel fraction were measured within three regions of interest. We controlled for cloud cover, snow cover and evaluated the impact of sensor radiometric differences between Landsat 7 and Landsat 8. Using Google Earth Engine to expedite data processing tasks, we show that there has been a weakly positive increase in the extent of subnival vegetation since 1993. Strongest and most significant trends were found in the height region of 5,000-5,500 m a.s.l. across the HKH extent: R²  = .302, Kendall's τ = 0.424, p < .05, but this varied regionally, with height, and according to the sensors included in the time series. Positive trends at lower elevations occurred on steeper slopes whilst at higher elevations, flatter areas exhibited stronger trends. We validated our findings using online photographs. Subnival ecological changes have likely impacted HKH carbon and water cycles with impacts on millions of people living downstream, but the strength and direction of impacts of vegetation expansion remain unknown.