Developing a High-Resolution, Gridded Data Product of Vapor Pressure Deficit Using Daymet and AmeriFlux

Regional and continental droughts are expected to become more frequent as a result of climate change-induced increases in surface temperatures around the globe. Vapor pressure deficit (VPD) is an important metric that quantifies atmospheric demand for water and can be used to assess drought conditions and evaluate water stress in plants. Increases in surface temperatures lead to higher atmospheric demand for water and higher values of VPD. High VPD can cause plant hydraulic failure, reducing photosynthetic gas exchange, and thus hindering vegetation's ability to uptake atmospheric carbon. Furthermore, high VPD can lead to vegetation drying down, and in turn, escalate the risk of wildfire. With expected increases in VPD, more research is needed to understand the differences in land-surface responses to elevated VPD across land cover and climate gradients. Thus, a fine-scale, gridded VPD product is needed to evaluate how land-atmosphere interactions are impacted by rising VPD. Here, we build two high-resolution (1 km, daily), gridded VPD datasets derived from Daymet and AmeriFlux for all of the continental US (CONUS) and parts of southern Canada and northern Mexico (CONUS+; i.e., the area between 25 and 50 degrees latitude and -125 and -67 degrees longitude) from 2001 to 2021. We use VPD measurements with land cover and climate information from 211 AmeriFlux towers to develop the Daymet-derived VPD for every pixel in the CONUS+ grid. VPD estimated from Daymet is validated against existing AmeriFlux tower records, NLDAS-2 derived VPD, and VPD computed from Daymet using a weighted temperature averaging scheme. We found that the land cover and climate VPD datasets perform similarly to one another when compared to AmeriFlux ground observations, with a root mean squared error of 0.13 kPa better for land cover and climate scaled VPD versus VPD computed using weighted temperature, suggesting that controlling for the influence of land cover and may improve daily estimates of VPD. Future users of these datasets will be able to study how VPD affects drought propagation and subsequently impacts ecosystem functioning, agricultural losses, and wildfire risk.