Bertrand Boeken, Yarden Oren and Moshe Shachak
Jacob Blaustein Institute for Desert Research
Ben-Gurion University of the Negev
Research questions
We investigated the effects of a severe drought year in 1999 with less than 50% of the long-term average rainfall, on annual plant communities in semiarid shrubland. The questions we addressed were how density and species richness of these communities responded to drought and how they start to recover after the drought. We studied these questions in two sites, in a field survey and in a manipulative field experiment.
Site description
The research site is situated in Sayeret Shaked Park near Ofakim in the Northern Negev Desert of Israel (31°17' N, 34°37' E). It is one in a network of Israeli LTER sites (Long Term Ecological Research). Rainfall, which only occurs in winter between November and March, has a long-term annual average of 200mm (Fig. 1). Average daily minimum winter temperatures are 6-8°C, and average daily maximum summer temperatures are 32-34°C. In the drought year of 1998/1999, total rainfall in Sayeret Shaked park LTER was 69 mm, amounting to 45% of the average over the last ten years (151 mm, Fig.1). It came after a rather rainy year with 177 mm (117%), and was followed by a relatively dry year with 105 mm (69%). The combination of a severe and less severe drought year in a row is rather rare (once in more than 60 yrs).
Fig. 1. Annual rainfall amounts in 1998-2000 in Sayeret Shaked Park, with long-term and 10-yr averages.
The research was done in semi-arid shrubland on two north-facing slopes with an angle of 5-10% in two watersheds. The soil is loessial, with 14% clay, 27% silt and 59% sand, and is at least 1.10m thick. Salt content of the 0-25cm soil layer is low, with electrical conductivity of 0.4 mMho. The surface is covered with a microphytic soil crust, consisting of bacteria, cyanobacteria, algae, mosses and lichens, and with scattered patches of the shrubs Noea mucronata and Atractylis serratuloides and Thymelea hirsuta.
Besides shrubs, the vegetation contains annual plant communities, which are responsible for most of the biomass production. Consequently, these communities are the main food source for livestock grazing in these shrublands. The annuals grow abundantly in the understorey of the shrub patches, and at significantly lower density on the crusted intershrub matrix. The annual plant communities consist of a few common or dominant species, and many less common to very rare species (of a total of close to 120 found in the watershed). The dominants (Stipa capensis, Bromus fasciculatus, Anagallis arvensis, among others) determine vegetation density, while the latter are responsible for species richness and diversity.
Field samples
We sampled annual plant communities non-destructively, by identifying and counting all plants in delineated pairs of a shrub and a crust patch, in both sites. In Site A, no manipulations were done on 32 pairs of samples of 20 x 30 cm per patch type. In Site B, 30 pairs of shrub and crust patches were sampled (sample area 0.17 and 0.33 m2, for shrub and crust respectively), 10 undisturbed and 20 disturbed. This entailed clipping of all aboveground vegetation and trampling of the soil surface during the previous years since 1995.
Results
In both datasets, plant density decreased from the very rainy year of 1998 to the drought year of 1999 (Fig. 2A and 3A). In the survey (Fig. 2A), it kept dropping in the low-rainfall year 2000, but not in the experiment (Fig. 3A), where plant density increased again from 1999 to 2000. Although shrub patches have higher density than crust patches, both show similar changes, in both studies.
Fig. 2. Plant density N, species number S and species richness R corrected for plant density in two patch types during three years.
In the survey species number dropped during the drought year of 1999 to half of the pre-drought values, but increased significantly during the next year, in spite of lower plant density (Fig. 2B). The variation in species number seems to follows that of the annual amount of rainfall. Both the drop in species number from 1998 to 1999 and its increase from 1999 to 2000 are more than expected based on density per se, so that species richness corrected for density decreases and than increases (Fig. 2C). This illustrates that the changes in plant density and species number are based on different groups of species, the dominants and the rarer/sparser species. The dominants and rarer species suffer similarly during the drought year, but the rarer species, though still with lower numbers than before the drought, recover faster.
Fig. 3. Plant density (A) and species richness (B) of annual plant communities in two patch types in controls and with trampling and clipping (10 pairs of patches each).
In the experiment at Site B, similar results were found (Fig. 3), except the rise in plant density during the last, dry post-drought year. It appears that the patterns of plant density and species richness both simply track rainfall, unlike the situation in the survey described above. However, examining the behavior of the individual species reveals that, as in the survey, the increase in species number and plant density from 1999 to 2000 was not due to recovery of dominant and common species, but of relatively rare species. Especially the most common species, the grass Stipa capensis, the small forb Plantago coronopus, and the minuscule grass Ammochloa palaestina decreased greatly in abundance. S. capensis was also most affected by trampling and clipping.
Discussion
We expected that the rarer species decrease more during a drought year and recover slower than common species. This is because rare species are strongly limited by site availability, especially during drought. Thus, populations of rare species become sparser or extinct. Seed production will be severely reduced, and thus recovery of the less common species was expected to be slow. Common or dominants, on the other hand, are generalists, able to grow in a variety of conditions. During rainy years (as 1998) their populations are very dense, and less likely to go extinct.
Our results indicate the opposite, since species number recovered rapidly but plant density kept declining in the post-drought year, also with low rainfall. We hypothesize that the discrepancy is caused by seed dormancy. Drought affected germination, growth and seed production of both common and rare species negatively. In rare species ungerminated seeds stay dormant, while common species rely predominantly on recent seed production. No seed reserves of common species were present after the drought in contrast to the rare species, which could respond rapidly to ameliorating environmental conditions. In addition, increasing species richness, and sometimes of plant density of rare species may be due to the absence of suppression or site pre-emption of weaker competitors by dominants, which suffered from the drought period.
Sede Boker, Feb 2001
Ben-Gurion University of the Negev