MAPPING AND PREDICTION OF DIEL AIR-WATER TEMPERATURE SIGNALS AT HIGH SPATIAL RESOLUTION IN A CATSKILLS, NY MOUNTAIN WATERSHED

Understanding heterogeneous diel stream temperature patterns throughout mountain watersheds is important for predicting the impact of climate change and ex-urbanization on cold water habitats. Daily temperature fluctuations directly affect aquatic organisms such as Salmonids, altering their physiology, behavior, and resilience. Recently, machine learning techniques (e.g., LSTM, MARS, and LSSVM) have been employed to predict daily average stream temperature characteristics, but spatially explicit water temperature forecasts at more ecologically relevant timescales (i.e., sub-daily) have remained elusive. The study of sub-daily stream temperature patterns has been often hindered by data quality and availability, as knowledge of site-specific geology, and groundwater/surface water connectivity influence prediction accuracy. The USGS has recently collected paired air-water temperature data at high spatial resolution (>60 locations) throughout the Neversink Watershed, Catskills, NY, where groundwater discharge plays a significant role in regulating water temperature in low order streams. Within the Neversink, diel air-water temperature signals have shown sensitivity to shifts in streamflow generation mechanisms, highlighting seasonal transitions (i.e., spring to summer) in local hillslope contributions from runoff to groundwater dominated flow paths. Consequently, these high-resolution observations enable the identification of stream reaches whose thermal budgets are significantly impacts by groundwater, and provide ‘groundwater aware’ metrics often missing from statistical stream temperature models. Standardizing water temperature using local air temperature along with other meteorological variables also allows for spatiotemporal trends assessment in landscape hydrologic connectivity in headwater stream networks. This study will use the high-resolution dataset for spatially forecasting of air-water temperature coupling in the form of established diel air-water metrics such as amplitude ratio, phase lag and mean ratio. Amplitude ratio is an appropriate metric for spatio-temporal groundwater discharge forecasting. Additionally, ML application in forecasting subdaily diel pattern of air-water temperature could be investigated in future studies.