Optimized NMME climate forecasts for rangeland management applications in the western United States

Authors: Schantz, Merilynn; Hardegree, Stuart; Sheley, Roger; ABATZOGLOU, JOHN - University Of California; HEGEWISCH, KATHERINE - University Of Idaho; Elias, Emile; JAMES, JEREMY - University Of California; Moffet, Corey

Submitted to: Rangeland Ecology and Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/15/2024
Publication Date: 4/17/2024
Citation: Schantz, M.C., Hardegree, S.P., Sheley, R.L., Abatzoglou, J., Hegewisch, K., Elias, E.H., James, J., Moffet, C. 2024. Optimized NMME climate forecasts for rangeland management applications in the western United States. Rangeland Ecology and Management. 94:207-214. https://doi.org/10.1016/j.rama.2024.03.008.
DOI: https://doi.org/10.1016/j.rama.2024.03.008

Interpretive Summary: Rangelands exhibit extremely high annual and seasonal variability in precipitation and temperature which in turn affects the phenology and biomass production of rangeland ecosystems. Seasonal climate forecasts would, therefore, be very useful to rangeland managers who typically have to make major management decisions far in advance of seasonal weather effects in any given year. Currently available seasonal forecasting tools are fairly coarse resolution and are not optimized for site-specific or regional forecasting applications. We used the site-specific seasonal forecasts from multiple different forecasting models to optimize predictions of monthly precipitation and temperature for 11 sites in 4 different ecoregions in the western US. We determined that site-specific optimization of forecasting produces higher skill predictions, and that multi-model ensemble forecasting usually produces higher skill forecasts than do individual models. We also determined that optimal model skill depends on the specific location, time of year that the forecast needs to be made, and the future time period for which the forecast is made. This allows for the opportunity to customize forecasts for specific locations and applications and produce higher skill seasonal forecasts than are currently available for generic applications. High skill seasonal forecasts could greatly improve the cost effectiveness of multiple rangeland management applications, as well as other agricultural and natural resource management applications across the US.

Technical Abstract: Weather plays a vital role driving rangeland management outcomes. While seasonal weather forecasting has been used by agricultural producers for decades, relatively few forecasting tools have been adapted for the diverse site and season-specific rangeland applications in the western United States (US). The development of seasonal climate forecasts at spatial and temporal scales useful to rangeland management provides new opportunities to inform climate sensitive activities, like restoration and forage production. We evaluated the utility of statistically downscaled precipitation and temperature climate forecasts derived from different seasonal forecast models at several rangeland sites in the western US. Forecast skill, defined by correlations between forecasts and historical observations, was generally better for temperature forecasts than precipitation forecasts with the highest overall skill seen in the Desert Southwest. We additionally used an optimization procedure that yields higher forecast skill than the traditional multi-model mean ensemble approach by using individual models or small-number ensembles. The combination of climate forecasts at actionable scales and model optimization elucidate significant forecasting skill at most locations that may reduce many of the prior barriers to incorporating such forecasts into management applications. While we demonstrate skill in climate forecasts, development and subsequent value of site-specific forecasting applications will require additional analyses to address stakeholder needs, human dimensions, the impact of weather variables on plant production, and economic impacts.