Using demographic modeling to develop post‐fire restoration strategies for a native shrub in a sage scrub community

Abstract

Mediterranean‐climate shrublands are key biodiversity hotspots and carbon storage pools, but are increasingly threatened by climate change, non‐native species, and altered fire regimes. Fires are important to historic shrubland disturbance cycles but can also promote non‐native plants, which may limit post‐fire native shrub recovery. Increasing drought with climate change could also reduce post‐fire shrub regeneration. I developed a stochastic, individual‐based demographic model (IBM) for the native shrub Artemisia californica, parameterized from an experimental removal of non‐native annuals after a 2013 fire in southern California. The IBM simulated A. californica recovery for 7 years after fire under different rainfall conditions (drought or pre‐drought) and non‐native removal strategies (from no years to all 7 years). Drought lowered A. californica canopy volume 7 years after fire by 90% or more. Rainfall in the second year after fire, when most A. californica germination occurred, had particularly strong effects on final canopy cover. Non‐native removal in all 7 years increased canopy volume by three times under drought conditions and 3.5 times under pre‐drought conditions. Targeting non‐native removal in the first 2 years proved nearly as effective, achieving from 88% (drought) to 95% (pre‐drought) the benefits of removal in all 7 years. In sum, low rainfall may be the most important limitation on post‐fire shrub recovery, but removal of non‐natives in years of pulsed shrub recruitment can be an effective restoration strategy even under drought conditions. More generally, this study illustrates how demographic models can help optimize the targeting of scarce management and restoration resources.