In water-limited ecosystems that are widespread across the western U.S., woody plant regeneration occurs infrequently and is influenced by complex environmental conditions that are often not well quantified. Despite the crucial role of regeneration for sustaining plant populations and communities, the influence of projected temperature increases and precipitation changes on regeneration remains poorly understood. Big sagebrush dominated ecosystems, widespread across the western U.S., are experiencing ongoing restoration challenges related to big sagebrush’s complex, episodic regeneration, and these challenges may increase in coming decades. Accurate assessments of long-term sagebrush regeneration outcomes can help design and improve resource management strategies. Here, we applied two models of big sagebrush regeneration, Schlaepfer2014 and Shriver2018, with different approaches and geographic extents to (i) quantify geographic patterns of big sagebrush regeneration probabilities for historical and future climate conditions, (ii) evaluate sensitivities of the models to sources of uncertainty in our simulation experiment, and (iii) contrast how well we can summarize model outcomes in terms of environmental site characteristics.
Results/Conclusions
Under 1980-2010 conditions, the process-based Schlaepfer2014-model predicted potential regeneration probability p = 0.44 across the entire distribution of big sagebrush ecosystems. The data-driven Shriver2018-model that represents big sagebrush restoration following fire and seeding, predicted p = 0.73 (across the Great Basin and the Snake River Plain) for the same 1980-2010 climate conditions. Simple environmental characteristics poorly summarized Schlaeper2014 outcomes, whereas a larger proportion of Shriver2018 was explained. The different model approaches became apparent in the Snake River Plain where Schlaepfer2014 predicted high potential regeneration, whereas Shriver2018 predicted low values reflecting current patterns of agriculture and cheatgrass invasion. Under end-of-century climate projections, Schlaepfer2014 projected average changes of +0.14; including increases for 52% of the ecosystem extent and decreases for 3% of the areas with >90% agreement among climate runs. Shriver2018 projected consistent decreases of 0.13 across 100% of the Great Basin and Snake River Plain. These two different models highlight the need for carefully designed approaches when projecting complex ecological processes: the process-based model suggested increased future potential regeneration across part of the big sagebrush range; however, the data-driven model suggested that active restoration management may become even more difficult.