Ecosystems are constantly perturbed by both natural and human-induced disturbances, ranging from storms and fires to grazing and clear-cutting. If an ecosystem has alternative stable states (under the same external conditions the system may be in more than one possible state), then a disturbance may push it from one state to the other, leading to a regime shift.
This phenomenon has been observed in various ecosystems, such as lake eutrophication due to fertilizer runoff and the bleaching of coral reefs. The system is then said to have undergone a catastrophic regime shift where recovery is either not possible or quite difficult to achieve, even when the initial disturbance is relatively weak.
This classic view of regime shifts, however, is often too simplistic. In many ecosystems where spatial processes play a significant role, such as savannas and drylands, regime shifts can take other forms due to spatial dynamics.
One reason for this is that disturbances are typically spatially localized so that a disturbed domain may be pushed to the alternative state, while the rest of the ecosystem remains intact. In a savanna landscape, this amounts to small-scale wildfires creating isolated patches of grasslands in an otherwise forested region. This will lead to a formation of a front, a region in space in which the properties of the ecosystem continuously change as one moves between one domain and the other. A few examples are shown in figure 1, showing fronts between different plant communities such as trees and grasses. The dynamics of such fronts are known in other fields such as physics and chemistry, but their behavior and implications in ecology have been mostly overlooked.
Fig. 1: Three examples of fronts between two states of vegetation in an ecosystem. From left to right, brush and forest in Glacier National Park (USA), pines and deciduous trees in Bohemian Paradise (Czech Republic), meadow and trees in Olympic National Park (USA). Photos taken by Yuval Zelnik.
In our paper, we decided to explore this issue and specifically look at the implications of front dynamics on the possible regime shifts that ecosystems may undergo. We focused on three main aspects of front dynamics: the behavior of a single front, the interaction between different fronts, and the possibility of more than one type of front under the same conditions. We used models that describe the dynamics of vegetation in drylands for this study, as such models have been successfully used to describe and explore the behavior of spatial vegetation patterns in drylands while being simple to analyze and understand.
The most straightforward consequence of front behavior on regime shifts is the velocity of the front, namely how fast a front moves and in which direction. Depending on the ecosystem in question, the physical and biological processes that control this property can be quite diverse. In drylands where lack of water is a strong stressor, relevant local processes include precipitation, evaporation and the growth of plants under water stress, while spatial processes include seed dispersal and water flow below and above ground. If overall water availability is high then the front will move to invade domains of bare soil, thus expanding the vegetation domains, while droughts will do the opposite, leading to a contraction of the vegetation domains.
If we change a given parameter, with all other conditions remaining the same, then there is one special value for which the front will not move at all. This is called the Maxwell point (named after the famous James Maxwell who revolutionized electro-magnetism), and calculating its value can allow one to predict in which direction a front will move. As we show in our paper, while knowing where the Maxwell point lies is useful, under highly variable conditions this may not be sufficient. For instance, it is possible that for a given average amount of annual precipitation, constant conditions will lead to the growth of the vegetation domain, while high variability between years will lead the vegetation domain to shrink. This will occur if the front will only move slowly forward in rainy years, while a drought of equal magnitude will lead to a fast contraction of the front.
The focus on the dynamics of a single front implies a simple relationship between front velocity and regime shifts. The direction of the front will determine what regime shift can occur due to a localized disturbance, while the speed at which the front moves determines how long it will take for this gradual regime shift to take place. This view neglects the possibility of front interactions, where two or more fronts affect each other’s behavior when they come closer to each other.
As we show in a simulation in figure 2, these front interactions can have a strong effect on the regime shifts associated with fronts. In particular, if fronts interact to slow each other down, as demonstrated in figure 2, then the fronts may not take over the entire system, and remnants of the alternative state may remain. This means that a desertification front, as might be caused by a shift in conditions such as global warming, need not lead to a completely barren landscape, but instead to small patches of vegetation, from which the ecosystem might be able to recover if conditions later improve. It also tends to lead to various spatial patterns of vegetation, as seen in the giraffe-like pattern in figure 2.
Fig. 2: Snapshots of a simulation showing an ongoing regime shift affected by front interactions. Higher biomass is shown by darker green colors, while white shows bare soil. Initially, a system with uniform vegetation is perturbed by a few small localized disturbances which create small domains of bare soil. Due to low precipitation, these bare soil domains slowly expand as fronts push into the vegetation domains. When the distance between fronts is small, they slow down due to front interactions, and eventually stop entirely, leading to green bands of vegetation. The specific pattern seen depends partially on the layout of the initial disturbances. Image courtesy Yuval R. Zelnik
Finally, it also possible that more than one type of front can exist under the same conditions. This situation is a little similar to the existence of alternative stable states, except that the front is a localized structure in space which tends to move and is thus not stationary. These different fronts have different properties, and we show how one front type may be significantly slower than the other. The existence of such front bistability means that one could slow down the progress of an ongoing regime shift by a limited intervention, focused on the region of the front itself. One could thus help maintain a viable ecosystem by a minimal change in the front region, instead of attempting to affect the entire ecosystem.
Overall, we show that different front properties can have significant ramifications to the fate of an ecosystem undergoing various disturbances. A better understanding of the ramifications of spatial structure and disturbances on ecosystems is important, since human impact, such as global warming and habitat fragmentation, is becoming more dominant.
These findings are described in a paper entitled Regime shifts by front dynamics, recently published in the journal Ecological Indicators. This work was conducted by Yuval R. Zelnik and Ehud Meron from the Ben-Gurion University of the Negev.