The enigma of terrestrial primary productivitySubmitted by editor on 30 January 2017.
Ecosystem productivity is extremely sensitive to small-scale variability of water and nutrient availability. Spatial variation in savanna vegetation in Okavango delta, Botswana. Photo credit: Petr Pokorny.
Winner of the E4 award
Irena Šímová and David Storch
Plants need water, sunlight and sufficient temperature to grow. It is thus no surprise that warm and wet tropical forests with year-long vegetation season are considered the most productive biome on the Earth. Tropical forests are also extremely diverse, productivity being often considered as a major factor limiting species richness. Nevertheless, scientists still disagree on the global productivity gradient, and there are actually many different models that predict the variation of productivity on Earth’s surface. We have reviewed different global productivity models and available sources of the field measurements of biomass production to reconcile different approaches to terrestrial productivity.
Productivity, i.e. the amount of biomass (or organic carbon) produced by plants per unit area and time, cannot be directly measured at large spatial scales. Instead, we need to rely on various models. Some of them simply assume that productivity is a function of temperature, precipitation or solar radiation. Other models are more complicated, including remotely sensed data of vegetation greenness or the dynamic response of vegetation under the changing climate. The models differ in their assumptions and input variables, and they consequently differ also in their outputs comprising geographic trends in productivity. This is somehow frustrating, given that productivity is a key ecological variable, which is often used in macroecological and global change studies. It is even possible that so far there has been no consensus concerning the relationship between productivity and species richness simply due to our inability to properly measure and estimate productivity.
Geographic patterns of annual NPP estimated by (A) the Potsdam model and (B) the MODIS-based model (averaged over years 2000-2012), and the residuals from the MODIS-based model when regressed against the Potstdam model (C).
Which global productivity model is closest to the reality? The most obvious way of model validation would comprise using the data of ecosystem productivity directly measured in the field. However, although both the field data and the majority of productivity models agree on the general decrease of productivity from the equator towards the poles, there is high discrepancy between modelled and measured values within individual climatic zones.
We outline three main reasons for this variability. First, field data are sampled at much finer scales than those used in models, so they do not represent random samples of the whole landscape. Second, besides water and sunlight, plants need nutrients in order to build their tissues. Whereas some models incorporate the effect of nitrogen availability on plant growth, the role of phosphorus, a key limiting nutrient in the tropics remains largely overlooked. Third, field data are rather heterogeneous, comprising a mix of natural vegetation and plantations.
Natural habitats and plantations show very different geographic trends of productivity. Natural stands (black circles) exhibit a linear decrease of aboveground net primary productivity (ANPP) with latitude whereas plantations or fertilized plots (white circles), by contrast, show a unimodal relationship, temperate areas reaching the highest values.
We thus strongly recommend future studies to sample productivity in the field by a standardized protocol, so that these data could be used to parameterize global productivity models. Such models then should combine all the major variables that limit plant growth, including nutrients. The era of Anthropocene is characterized by fast rates of habitat degradation and human-induced climate changes. Productivity is a variable of primary interest as it is related both to resource availability, potentially limiting biological diversity, and to the dynamics of the carbon cycle. A clear consensus on global productivity patterns would enable us to understand the drivers of biodiversity variation, and to make better predictions of vegetation and biodiversity changes in the near future.