Long‐term mortality patterns of the deep‐rooted Acacia erioloba: The middle class shall die!

Abstract

Question: Is there a relationship between size and death in the long‐lived, deep‐rooted tree, Acacia erioloba, in a semi‐arid savanna? What is the size‐class distribution of A. erioloba mortality? Does the mortality distribution differ from total tree size distribution? Does A. erioloba mortality distribution match the mortality distributions recorded thus far in other environments? Location: Dronfield Ranch, near Kimberley, Kalahari, South Africa. Methods: A combination of aerial photographs and a satellite image covering 61 year was used to provide long‐term spatial data on mortality. We used aerial photographs of the study area from 1940, 1964, 1984, 1993 and a satellite image from 2001 to follow three plots covering 510 ha. We were able to identify and individually follow ca. 3000 individual trees from 1940 till 2001. Results: The total number of trees increased over time. No relationship between total number of trees and mean tree size was detected. There were no trends over time in total number of deaths per plot or in size distributions of dead trees. Kolmogorov‐Smirnov tests showed no differences in size class distributions for living trees through time. The size distribution of dead trees was significantly different from the size distribution of all trees present on the plots. Overall, the number of dead trees was low in small size classes, reached a peak value when canopy area was 20 ‐ 30 m2, and declined in larger size‐classes. Mortality as a ratio of dead vs. total trees peaked at intermediate canopy sizes too. Conclusion: A. erioloba mortality was size‐dependent, peaking at intermediate sizes. The mortality distribution differs from all other tree mortality distributions recorded thus far. We suggest that a possible mechanism for this unusual mortality distribution is intraspecific competition for water in this semi‐arid environment.

Estimating tree abundance from remotely sensed imagery in semi-arid and arid environments: bringing small trees to the light

“…Thus the choice of the negative exponential in our case was well justified. However, the case study presented in this paper is based on one of the largest tree databases and probably the largest tree database in the African continent where data are time and space replicated (Moustakas et al 2006). In most cases the remotely sensed derived number of tree individuals is lower and thus the accuracy is expected to be lower.…”

“…As scientists studying paleontology, paleoecology and climatic change seek long term data to draw sound conclusions about possible trends in ecosystems (e.g. Weiss et al 1993), applications of remote sensing are becoming more popular and often provide the desired long-term data (Clark et al 2004;Moustakas et al 2006). Though tree abundance derived from remote sensing is a good index of the current situation, the seedlings and saplings, (i.e.…”

“…Furthermore, tree mortality in general is also size dependent (e.g. Lorimer et al 2001;Moustakas et al 2006), and therefore the death of a tree depends usually on its size (for a detailed review see Harcombe 1987). Thus, in order to evaluate the accuracy of the proposed method one needs to know the mortality and growth size class distributions.…”

“…We apply the proposed method to a spatio-temporal tree database derived with a combination of remote sensing methods and field work (for further details see, Guenther 2002;Moustakas et al 2006). In the present study, we use data pertaining to four different years, 1964, 1984, 1993 and 2001 and five mixed-species plots that contain around 14,500 tree individuals in total.…”

“…The projected canopy surface area of each tree can be easily extracted using remote-sensing techniques, at least in the case of open forest canopies (e.g. Clark et al 2004;Moustakas et al 2006). Thus, the parameters of the negative exponential size-class distribution in the study area can be readily estimated.…”