Using AI technology to predict Snow Water Equivalent for western U.S.
Accurately estimating precipitation quantities and understanding water flow is crucial for diverse applications, including disaster forecasting and agriculture. A significant aspect of this challenge lies in estimating Snow Water Equivalents (SWEs), which play a vital role in water flow dynamics. Unfortunately, accessing this data has been daunting for “laypeople.” Moreover, climate change is rapidly impacting hydrological models, rendering them inaccurate and demanding innovative replacements. To address these issues, this paper introduces a series of beginner-friendly machine learning (ML)-based models designed to leverage historical SWEdata and predict future values at any location. In our research, we collected real-time climate data from various climate monitoring satellite projects such as Sentinel, MODIS, and model datasets like gridMET, terrain data, alongside historical SNOTEL data from ground-based measurements. This diverse dataset served as the foundation for training and testing a range of state-of-the-art ML models. These models encompass a variety of approaches, including Random Forests, XGBoost, and Long Short-Term Memory (LSTM) networks. By employing these cutting-edge ML algorithms, we aim to make the estimation of SWEs more accessible and comprehensible to a wider audience. The incorporation of diverse datasets from satellite projects and ground-based measurements ensures robustness and accuracy in our predictions. Through this research, we strive to create a user-friendly platform for individualsand experts alike to access valuable SWE information, enabling better-informed decisions in water resource management, disaster preparedness, and other critical domains. The potential impacts of this work extend far beyond its immediate applications, contributing to the ongoing battle against the adverse effects of climate change. By equipping users with the capability to accurately estimate SWEs, we can better anticipate water availability and its consequences in the face of a changing climate. Ultimately, this research represents a significant step forward in harnessing the power of machine learning for addressing pressing environmental challenges and promoting sustainable resource management practices.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.