Impacts of Biodiversity Loss

See allHide authors and affiliations

Science  04 May 2012:
Vol. 336, Issue 6081, pp. 552-553
DOI: 10.1126/science.1222102

Historically, ecologists and evolutionary biologists have treated the variety of life on Earth as if it were a simple by-product of the physical and chemical variation that generates biological diversity and allows it to persist. However, this perspective changed in the 1990s, when scientists began to manipulate biodiversity in controlled environments and found that it can act as an independent variable that directly controls ecosystem-level functions, such as nutrient cycling and biomass production (14). The idea that biodiversity might control—rather than just respond to—Earth's biophysical processes was foreign to many researchers (5). But by 2010, more than 600 manipulative experiments had been performed, spanning much of the tree of life and most major biomes on the planet (6). We now know that biodiversity regulates many ecosystem-level processes, including some that are essential for providing goods and services to humanity (69). On page 589 of this issue, Reich et al. (10) provide important novel insights into how much diversity is needed to maintain the productivity of ecosystems.

The authors reanalyze data from two classic biodiversity studies that have been running for more than a decade at the Cedar Creek Ecosystem Science Reserve in Minnesota. By fitting data collected over a 15-year period to several mathematical functions (linear, log, power, and hyperbolic), the authors quantify the form of the relationship that describes how plant species richness influences the production of plant biomass. They show that the effects of biodiversity on productivity change from saturating functions that are prominent early in the experiments (see the figure, panel A), to monotonically increasing functions later in the experiments (panel B).

Scaling diversity-function relationships.

Since 1990, more than 600 experiments have manipulated the diversity of plants, animals, fungi, protozoa, and bacteria in a variety of Earth's biomes. These studies have shown that ecosystem functions like nutrient cycling and biomass production are positively related to biodiversity, but that relationships saturate at relatively low levels of diversity (A). Reich et al. have reanalyzed results from two long-term studies of grassland plants and found that although saturating functions are prominent early in the studies, diversity-function relationships ultimately become monotonically increasing given enough time (B). Short-term experiments may thus underestimate the number of species needed to maintain ecosystem-level processes. If the results prove to be general, Reich et al. will have quantified how the ecological impacts of extinction scale through time (A to B). If others can similarly quantify how diversity-function relationships change with the spatial extent of studies (A to C), we would have scaling relationships to estimate the fraction of species needed to maintain ecological processes in more realistic ecosystems (D).


Reich et al. argue that the reason for this change is that it takes time for species to express the biological traits that allow them to fill their various ecological niches. They present a set of calculations that estimate how much of the diversity effect in any given year is driven by processes involving two or more species (called complementarity). They show that complementarity grows stronger through time, and this trend is associated with a greater divergence in the biological traits of species in the experimental plots. These trends are not conclusive evidence that niche differences are the underlying cause of the reported patterns, but they hint at the possibility that biological “niche space” becomes more completely filled as communities interact and assemble through time.

Several studies have shown that diversity effects grow stronger with time (11, 12), but Reich et al. go further by quantifying how the shape of the diversity-function relationship—which tells us what fraction of species is required to maintain ecosystem functions—changes through time. If biodiversity has a saturating effect on ecosystem processes, as most prior studies suggest, this implies that some fraction of species are functionally “redundant,” and can be lost with little or no impact on ecosystem processes. Ehrlich and Ehrlich (13) compared biological redundancy to the redundancy of rivets on an airplane wing. Loss of one or few rivets will not affect the performance of the plane, because wings are engineered with an excess of rivets. But lose one too many rivets, and the loss could have catastrophic consequences for passengers on the plane.

If, however, the relationship between biodiversity and ecosystem functioning is monotonically increasing, as Reich et al.'s reanalyses suggest, then each extinction would produce an incremental decrease in the functioning of ecosystems. This scenario would be far more pressing for conservation. A notable fraction of Earth's biodiversity has already been lost, and given current rates of extinction, much more is likely to be lost in the coming century (14). If the results of Reich et al. hold generally true, then biodiversity loss has probably already begun to degrade essential processes that sustain the productivity of ecosystems.

Reich et al.'s findings emphasize the importance of long-term studies, such as those sponsored by the U.S. National Science Foundation's Long Term Ecological Research Program that partially funded the Cedar Creek experiments. Most natural communities do not develop on the 1- to 3-year time frame of a typical grant, which is the scale of the average experiment. It is, therefore, possible that many biodiversity experiments to date have revealed just the tip of the iceberg—the short-term, transient effects of biodiversity on ecosystem processes. The real impacts of diversity loss could be much greater.

It remains to be shown whether the results of Reich et al.'s study are general, or whether something unique about the species pool, environmental conditions, or experimental methodologies make the experiments at Cedar Creek the exception rather than the rule. Resolving this will require comparison to other long-term studies, and reanalysis of data from biodiversity experiments involving even longer population dynamics than those at Cedar Creek (such as studies performed with model systems of bacteria or algae).

But if the conclusions of Reich et al. hold generally true, and monotonically increasing diversity-function relationships are indeed the norm, then this study moves us an important step closer to predicting the ecological consequences of diversity loss in real ecosystems, where life forms have evolved and interacted for many generations (see the figure, panel B). The study should also stimulate others to ask similar questions about how the form of diversity-function relationships changes with the spatial scale of experiments (see the figure, panel C). Once we know how diversity-function relationships scale in both time and space, we will have the statistical models needed to forecast the ecological consequences of extinction from whole ecosystems. For a field of research that did not even exist until the 1990s, development of such models would represent monumental progress in a remarkably short time.


Navigate This Article