Cook Agronomy Farm Snow Coverage Data Analysis

In the Pacific Northwest, climate change is expected to cause decreases in summer and fall precipitation and increases in growing-season evapotranspiration. Because many fields in the region use dryland cropping, the crops are heavily dependent upon the water that is held in the soil over the summer that was deposited during the wet months. Since soil water content is dependent upon snow cover, it is important to understand and quantify snow conditions so that farmers can make decisions regarding farm management to maximize crop efficiency, sustainability, and health. The snow conditions on the Cook Agronomy Farm in Pullman, Washington were analyzed using Phenocam image data gathered during the 2017-2018 winter. A systematic method to characterize the snow conditions was developed, tested, and applied by analyzing the data using script written in Python. Using images from a digital time-lapse camera co-located with an eddy covariance tower, we developed a simple method to quantify the percent of snow cover present within each image. The color indices of the images were analyzed from 10:00 to 14:00 local time over the period of a week, and the ratios of amount of blue to amounts of red and green in the images were used to identify snow cover. Based on a visual comparison of the modeled snow coverage with the observed snow cover, this method of quantifying snow works with a high degree of accuracy under opportune conditions; however, broader testing of the method indicates that additional refinement is required to exclude erroneous snow identifications. The resultant method identifies the presence of snow cover and quantifies it as a percentage of snow cover in the image. Trends in percent snow cover and other variables such as soil water content, net radiation, and evapotranspiration were compared. Quantification of snow coverage has potential to be used for field-scale hydrologic models and provides valuable insight to the role of snow in the water cycle.