Köchy, M., and K. Tielbörger (2003)
We used a spatially-explicit individual-based model to simulate the effect of mean annual precipitation level (maP; low: 75 mm, medium: 90 mm, high: 105 mm) and the shape of interannual distribution of precipitation (more normal vs. right-skewed) on the performance of annual plant populations. Our model simulates soil moisture, seed germination, and plant growth calibrated to the desert annual Filago desertorum (Asteraceae), assuming a homogenous environment located in the northern Negev desert (Sede Boqer, Israel). We used two approaches for producing rain scenarios. (1) We increased or decreased daily rain values of Sede Boqer to change the shape of the maP distribution without changing the intraannual rain pattern. (2) We reproduced the same maP distributions (shape and level) by selecting years of data from several Israeli climate stations with different intraannual rain patterns. We compared population performances for 20 simulations of 100 years for each scenario with an index based on the percentage of years when seed production was greater than the 75-percentile under unmanipulated 1961-1990 Sede Boqer precipitation and subtracted the percentage of years when seed production was less than the 25-percentile. Factorial ANOVA showed that performance differed significantly among approaches, maP levels, and distribution shape. Performance always increased with maP (low: -0.31, medium: -0.07, high: 0.14). Performance at high maP was significantly higher under normal than under skewed distribution for both approaches. This was reversed when maP was low, but the difference was significant only for approach 1. Performance at medium maP was significantly higher under normal distribution for approach 1 but did not differ between distributions under approach 2. The results show that, subject to model assumptions, population performance varies not only with maP but also with the shape of annual precipitation distribution and intraannual variation of precipitation.