Long-term spatial-temporal trends in saguaro recruitment: local buffering in the response to environmental conditions

Background/Question/Methods

Contemporary climate in the Sonoran Desert is characterized by an increased frequency of extreme temperature events, which may reduce the recruitment of perennial species.  Maintaining the ecological integrity of this desert requires a mechanistic understanding behind population-level recruitment variation at different spatial-temporal scales, particularly with iconic species such as the saguaro (Carnegiea gigantea). Large to medium scale changes in recruitment conditions may not occur locally, and long-term population persistence might benefit more from local variability in recruitment success than from larger-scale responses. Here, we test whether local changes in recruitment events are decoupled from changes in environmental conditions.  We investigated spatial-temporal variation in successful recruitment events of a saguaro population using spatially-explicit records of nearly 5800 individuals monitored since 1964 at four-11 Ha, permanent plots in Tumamoc Hill; in Tucson, AZ. We used quantile regressions with cubic splines to estimate year of recruitment for all individuals. We incorporate variability at three different levels: at the local level (20 x 20 quadrats) by fitting 25^th , 50^th, and 75^th quantile regressions of the inverse of annual growth rates versus height (m); at the plot level by fitting separate curves for the plots that were set up along each slope aspect of Tumamoc Hill (north, south, west, and east); and at the population level by incorporating long-term changes in growing conditions since the 1700’s. We used Kendall’s Tau at the site, plot, and population levels to test for 1) population level response to climatic conditions depicted by Palmer’s drought severity index (PDSI), 2) synchronic peaks of recruitments among plots, and 3) consistency in recruitment sites within each plot.

Results/Conclusions

Our results indicate a weak (but significant), population-level positive correlation between recruitment peaks and favorable recruitment conditions (high PDSI), especially between 1900 and 1940 (0.11< Tau<0.23). We detected a lack of synchronicity between plots with more favorable conditions (south and east had recruitment below expectations between 1970 and 2012) and plots with less favorable conditions (west and north had recruitment above expectations), especially in the last 25 years.  Within each plot, there was an overall high spatial correspondence (Tau > 0.5 at 20x20 m scale) among sites of successful recruitment from period to period regardless of the environmental conditions. Despite a substantial reduction in recruitment in the last 25 years, local conditions buffered recruitment from larger scale changes in environmental conditions, which might be a direct consequence of location and abundance benefactors and competitors.