Abstract ID: 3.9087 | Accepted as Talk | Talk/Oral | TBA | TBA

The snow dominated headwaters in the mountains of the Western United States (US) are an essential seasonal water resource. Predicting the spring release timing and magnitude requires accurate spatial information of snow distribution and extent during the accumulation and ablation seasons. One way of trying to capture the accumulation phase of seasonal snow is through the use of spatially distributed snow models that simulate the snowpack from meteorological observations. One of the dominant inputs that dictates a model’s ability to accurately simulate snow cover is precipitation, which to date is either recorded through sparse in-situ measurements stations or modelled with numerical weather prediction models (NWP). The NWP models are initialized with spatial or point observations that often have limited coverage in mountain areas, leading to underestimation or missing areas of snowfall. Approaches to improve the quality of precipitation input data for snow models include the use of spatial snow depth distribution patterns to inform precipitation/snowfall estimation, correcting the homogeneous precipitation outputs of standard interpolation methods with measured accumulation patterns. The number of spatial snow depth observations have increased over the recent years, with lidar-based platforms dominating the underlying technology. In this work, we evaluate the use of snow distribution maps to inform the precipitation input of a spatially distributed snow energy balance model to improve simulated snow depth accuracy. The lidar-based depth maps were recorded over multiple winter seasons (2020 – 2024) in Mores Creek, Idaho, US, and capture a range of accumulation seasons from below to above average years. Normalization across flights during the accumulation period showed a promising potential to derive precipitation information that can be used across water years to reduce the need for repeated observations and enhance the scalability of this approach.