Predicting Playa Inundation Using a Long Short‐Term Memory Neural Network

• Published in: Water resources research 2021-12, Vol.57 (12), p.n/a
• Main Authors: Solvik, Kylen, Bartuszevige, Anne M., Bogaerts, Meghan, Joseph, Maxwell B.
• Format: Article
• Language: English
• Subjects: Agriculture; Aquatic habitats; Aquifers; Birds; Climate; Climate change; Conservation; Drought; Drying; Evaporation; Groundwater; Habitats; High plains; Hydrologic models; Hydrology; Lakes; Land conservation; Land use; Learning algorithms; Machine learning; Migratory birds; Migratory species; Modelling; Neural networks; Playas; Probability theory; Rain; Rainstorms; Recharge; Storms; Water shortages; Wetlands
• ISSN: 0043-1397; 1944-7973
• DOI: 10.1029/2020WR029009

In the Great Plains, playas are critical wetland habitats for migratory birds and a source of recharge for the agriculturally important High Plains aquifer. The temporary wetlands exhibit complex hydrology, filling rapidly via local rain storms and then drying through evaporation and groundwater infiltration. Using a long short‐term memory (LSTM) neural network to account for these complex processes, we modeled the probability of playa inundation for 71,842 playas in the Great Plains from 1984 to 2018. At the level of individual playas, the model achieved an F1‐score of 0.522 on a withheld test set, displaying the ability to predict complex inundation patterns. When simulating playa inundation over the entire region, the model is able to very closely track inundation trends, even during periods of drought. Our results demonstrate potential for using LSTMs to model complex hydrological dynamics. Our modeling approach could be used to model playa inundation into the future under different climate scenarios to better understand how wetland habitats and groundwater will be impacted by changing climate. Plain Language Summary Playas are small, rain‐fed lakes typically found in the Great Plains of the US. Most of the time, they are dry, but when filled they provide important wetland habitat for migrating birds. As the water drains from the playas, they help to recharge the High Plains aquifer, which provides much of the water for agriculture in the region. We used a machine learning model to predict when individual playas are wet and when they are dry using weather, playa size, and information about the land use adjacent to each playa. Our model can accurately predict when playas fill and drain, valuable information for conservation efforts in the region. This research can be used by conservation managers and land‐owners to help protect these critical wetlands. Key Points Playas are infrequently inundated critical wetland habitats for migratory birds and a source of recharge for the High Plains aquifer Modeling playa inundation is challenging but highly valuable for conservation efforts, especially under a changing climate Using a convolutional neural network, we can accurately predict monthly inundation for 71,842 playas across the Great Plains