Although theoretical models predict that drylands can experience sudden shifts, empirical evidence on this topic is very scarce and shows contrasting results. For instance, while Gao et al. (2011) found a degradation threshold (≈20% vegetation cover) for natural restoration of overgrazed rangelands in a long-term (35-years) observational study in China, Bestelmeyer et al. (2013) found no critical thresholds, even at low plant cover values, in Chihuahuan Desert grasslands after a long term (13-year) pulse-perturbation experiment of heavy grazing and shrub removal. Different strengths and modulating factors of the ecohydrological feedbacks studied in CASCADE could provide insights on these apparently contradictory results. Although observational field experiments are essential to illustrate sudden shifts in ecosystems, they are often unable to provide conclusive results, which is in part due to the background environmental heterogeneity of the landscape, particularly strong in drylands, but mostly to the complex interactions occurring between multiple control factors. Conversely, appropriate manipulative experiments allow disentangling the relative role of the factors involved, yet they may imply an over-simplification of real ecosystems. CASCADE has adopted a combination of mesocosm and field manipulative experiments with field observations as the most promising approach for the study of the feedback mechanisms that may trigger sudden shifts in ecosystems.

In order to disentangle the various components of the ecohydrological feedbacks that relate plant pattern, resource availability and productivity in drylands, as well as the independent role of critical factors that control these feedbacks, CASCADE has followed a fully manipulative experimental approach combined with field observations. Manipulative experiments allow the isolation of the processes and factors of interest, thereby facilitating the understanding of the underlying mechanisms and providing useful information for developing, parameterizing, calibrating and validating general models. Three interlinked experiments were used to determine:

1. The independent role of plant cover and plant pattern on resource (water, soil) conservation in drylands.
2. The two-way feedbacks between plant pattern and resource conservation, and the role played by plant diversity in modulating these feedbacks.
3. The relevance of local transfer of resources from bare-soil inter-patches to downslope plant patches for plant performance and patch productivity.

The first two experiments were conducted on a set of 24 closed (2 x 1 m) plots, which allowed event-based monitoring of runoff and sediment yields, and where patch cover and pattern were manipulated in order to create a variety of patch spatial patterns.

• The first experiment »Spatial pattern and resource redistribution in drylands focused on assessing one side of the feedback process: pattern→resource conservation. By using inert materials, we mimicked the structural role of vegetation on resource conservation, but avoided the potential response of vegetation to the resulting changes in resource availability; this way the feedback loop was artificially broken, facilitating the independent assessment of one of the components.
• The second experiment »Feedbacks between plant pattern and resource conservation used real plant communities, artificially arranged in patterns of interest; this way both sides of the feedback loop (the effect of pattern on resource conservation and the effect of resource availability back to vegetation) were assessed.
• The third experiment »Patch-scale effect of resource redistribution on plant performance focused on further assessing the local aspect of pattern-resource feedbacks (i.e., the effect of local redistribution of resources, from inter-patches to downslope patches, on plant performance) in natural communities, and used natural slopes from a degraded semiarid area on which a variety of shrub seedlings were planted in 2004.

Synthesis of results

By performing these manipulative experiments and observations, we have tested the main ecohydrological mechanisms and processes underlying the hypothesized feedback loops that drive dryland dynamics, response to stress, and potential sudden shifts in drylands.

Our findings have demonstrated that both plant cover and plant pattern exert a critical role in controlling water and soil conservation in patchy ecosystems. This role mainly relies on the sink capacity of the soils underneath the plant patches, rather than on the capacity of the patches for rainfall interception and physical obstruction to overland flow. The connectivity of bare-soil emerged as the most critical pattern attribute for explaining the hydrological behavior of patchy ecosystems, as it reflects and depends on both cover and pattern. Larger bare-soil connectivity implies larger water and sediment losses from semiarid slopes, but it also implies larger inter-patch areas, which is beneficial for the performance of the downslope patch.

Our results provide critical insights on the control factors of source–sink dynamics in semiarid lands. Spatially explicit or mechanicist models that investigate the interactions between spatial vegetation pattern and resource redistribution (e.g. Urgeghe et al., 2010; Mayor et al. 2013), as well as ecogeomorphic evolution models (e.g. Saco and Moreno-de las Heras 2013) may greatly benefit from the empirical findings presented here.

Understanding the control factors that drive plant performance and ecosystem productivity in semiarid lands is critical to the conservation, management and restoration of these areas. The evidence for a positive relationship between seedling growth and the size of the upslope inter-patch area should be considered when designing conservation and restoration actions in semiarid lands. Along these lines, treatments that exploit and enhance source–sink dynamics on dryland slopes can improve the re-introduction of native shrubs into areas under strong water–stress conditions. Furthermore, with the aim of recovering previous landscape processes and minimizing resource leaks, the spatial pattern of the introduced seedlings should pursue a functional patchiness and source:sink area ratio, that maximizes both vegetation cover and the amount of water input that can be captured by the vegetation. Further research is needed to better define this optimum source:sink ratio for a number of plant communities and spatial scales.

Finally, although plant cover and biomass are the most common vegetation properties used for hydrological modeling, our results suggest that other patch metrics like patch number and/or size distribution could be better hydrological indicators than patch cover. Integrated indexes based on capturing the connectivity of the bare-soil matrix in patchy ecosystems, such as Flowlength index, have great potential as surrogates for the hydrologic functioning in semiarid landscapes. These indices can be easily obtained from aerial photographs and incorporated into hydrologic and erosion models at the hillslope and catchment scales.

Overall, the results reported here support the hypothesized role of ecohydrological processes and feedbacks as potential inside mechanisms underlying sudden shifts in drylands. Further experiments will test how increased pressure on dryland systems could trigger sudden shifts towards degraded states, and how this degradation can be reverted by manipulating plant cover and diversity.

»References