Where are ecosystems losing resilience?

In 1992 the North Atlantic cod fishery collapsed, from a productivity of over 200 thousand tons per year to less than 1% of the historical record. Millions of dollars and more than 40 000 jobs were lost. Today, about 35% of all fish stocks are currently over fished, at risk of similar impacts¹. The global oceans oxygen network has identified >500 sites where hypoxia is occurring. Low oxygen zones can cause fish die-off events and toxic waters for human health^2,3. Meantime, over 1300 forest plots have been documented to undergo die-off events⁴, and some areas of the Amazon rain forest have already been observed to shift from carbon source to carbon sinks⁵, losing the forest ability to regulate climate. Despite being studied in different scientific disciplines, these phenomena have some features in common: they are large, abrupt and persistent changes in the function and structure of systems commonly known as regime shifts⁶. They are caused by nonlinear dynamics (tipping points), and they impact the benefits people gets from nature. Their nonlinearities make them hard to predict, but most importantly, difficult to impossible to reverse – at least on the time scale relevant to managers and governments.

As a MSc and PhD student I became interested in regime shifts, their common causes and potential impacts on ecosystem services and human wellbeing. At the time we did not know how common they are, or their main causes and impacts. We developed the regime shifts database to better understand their diversity and assess common patterns or sources of uncertainty⁶. We found that climate change is certainly a common cause, but many other pressures are at play. Ignoring them can undermine our ability to successfully avoid or manage regime shifts. Other drivers are related to food production (fishing, agriculture, fertilizers use), urbanization, and disturbing the water cycle^7,8. While many impacts on ecosystem services where expected, cultural services were surprisingly up the rank highlighting an aspect that ecologist often ignore on their studies⁹.

A fundamental yet elusive question is whether regime shifts can be interconnected. If the occurrence of one regime shift increases the probability of another ecosystem tipping over, it means they cannot be managed assuming they are independent⁸. If they are connected, the patterns of connections could redraw the geopolitical map of the world. It will make explicit how a nation failing in managing their resources could impact the ability of another nations to benefit from their natural capital. For example, if Brazil allows the collapse of the Amazon rain forest and parts of the ecosystem tips towards savannas, the rainfall contributions from the Amazon to the Andes region will be reduced, limiting agricultural productivity in countries like Colombia, Peru or Ecuador where agriculture is the main engine of their economy. But how could ecosystems being connected, what are the mechanisms or the variables to look for? It turns out that by looking at regime shifts as causal networks we can design algorithms that explore a variety of plausible causal mechanisms⁸. We found that mechanisms responsible for cascading effects are often related to modifications in nutrients and water cycles, agriculture, soil erosion, changes in fire frequency and intensity, as well as climate related processes (temperature, precipitation, extreme events). While we partially solve the problem of how, we have not yet understood if these mechanisms are strong enough to couple far away places. The empirical evidence is still missing, it remains an open problem.

To solve it, we first need to understand where on Earth are regime shifts likely to occur? The answer will give us a series of candidate places where we can test causal connections, but finding them depends on our ability to observe and measure resilience¹⁰. Resilience is the ability of any system to withstand change while still keeping its structure, functions and hence its identity. In short, it is the capacity of a forest to still be a forest despite droughts, fires, or deforestation; or the capacity of a coral reef to recover from disturbances such as heatwaves. As ecosystems loss resilience, they leave statistical signatures in time series that can be interpreted as early warnings of regime shifts¹¹. Using remote sensing data of primary productivity for marine and terrestrial ecosystems, I computed different early warning statistics seeking signals of resilience loss.

Around 29% of the terrestrial ecosystems and 24% of marine ones are showing symptoms of resilience loss¹⁰. Some of the most affected areas include the boreal forests and taigas in Canada and Russia, as well as the Amazon and South East Asia rain forests. These places have independently being identified as risk areas where temperature thresholds can shift primary productivity from carbon sink to carbon source, potentially degrading their role in regulating climate⁵. The spatial patterns also align with studies in drylands, where scientist estimate that 28.6% of drylands world wide could cross thresholds by 2100 in some climate models¹². Strong symptoms of resilience loss are also found in areas of the ocean where upwelling systems are dominant, for instance the California current, Canary Islands or the Benguela systems. Upwellings are ocean currents that bring nutrient rich waters from the bottom of the sea up to the surface. Too little nutrients can decrease primary productivity, too much can produce coastal eutrophication and hypoxia (low oxygen zones)¹³. The most productive fisheries of the world are located in upwelling systems.

Imagine if we could observe and assess ecological resilience in real time. We would be able to foresee some regime shifts, prepare for those we cannot prevent, or implement preemptive managerial actions for those we can. In fact, some of my ongoing work tries to identify companies and financial actors exposed to the risk of tipping points, and with the agency to do something about it^14,15. We are close to develop a resilience observatory, a tool that can empower nations, cities or businesses to take into account ecological vulnerabilities in their planning and policies.

Ecosystems are showing symptoms of resilience loss A global analysis of primary productivity in marine and terrestrial ecosystems shows that about a third of the areas analyzed are already showing early warnings of regime shifts. Orange dots showcase about 3000 case studies of regime shifts already observed based on literature review. Source: the regime shifts database