Ecosystems in four dimensions

The ecosystem concept is perhaps the most important concept in ecology for scientists (Cherrett, 1989) and the public at large. Eighty years ago, Arthur Tansley (1935) noted that the method of science entailed separating parts of the universe for intensive study. These isolated parts, or systems, are included in larger systems that overlap, interlock, and interact with one another. The isolation of the parts is artificial in many ways, but necessary for advancing understanding. Sixty years later, Reiners (1986) asserted that ecosystems were dimensionless, ranging in scale and scope from a single decaying acorn to the entire planet's biosphere. But if an ecosystem can be anything from an acorn to a planet, how useful can this concept be for scientific inquiry? The power of the idea depends on how the boundaries of ecosystems are defined and addressed. The spatial extent of ecosystems has often been delimited by abrupt structural changes, such as the boundaries of a lake, and the border of a forest and a meadow. Ecosystems are also demarcated within a single type of vegetation, such as the boundary between forests of different ages, or different dominant plant species. The classical importance of defining spatial boundaries became less critical with the development of geographical information systems and mathematical approaches for dealing with gradients of change without requiring sharp boundaries or covarying changes in ecological factors. Modern conceptualization of ecosystems may embrace the sorts of overlapping, interlocking, and interacting factors noted by Tansley, allowing a more diffuse consideration of the areal bounds of any system of interest. Whether the defined boundaries of an ecosystem are sharp or diffuse, it is important that the boundaries do not leave out important processes and components. A forest ecosystem that was defined to contain only the aboveground portions of trees would not be a viable object of study for hydrology, physiology or biogeochemistry. In this issue of New Phytologist, Richter & Billings (pp. 900–912) discuss the critical need to include depth and timescale of ecological processes below the soil surface, and demonstrate the value of extending classic thinking about ecosystems to include the full ‘critical zone’ concept from Earth science. ‘But if an ecosystem can be anything from an acorn to a planet, how useful can this concept be for scientific inquiry?’ A classic example of the importance of the critical zone issues of depth and time on ecosystems was described by the soil scientist, Hans Jenny, for a sequence of forests on the coast of California (Jenny et al., 1969). The ‘ecological stairstep’ consists of a consistent parent material for soil development, with varying lengths of time for ecological processes to develop (Fig. 1). Terraces raised above the ocean for 100 000 years are dominated by grass-like vegetation, with a strongly humified upper mineral soil. After another 100 000 years of ecological processes, the terraces are dominated by redwood and Douglas-fir trees; the upper soils have begun to loose iron and other elements, leaving quartz crystals behind. Soils and vegetation continue to change over millennia, with shifting dominance toward tall Bishop pine trees as the soils continue to weather and soil horizons deepen. The terrace locations closer to the terrace edge have better drainage of water during wet winters. Well-drained locations have had the mobilized iron moved farther down the soil profile. The more poorly drained locations to the rear of the terraces have accumulated a dense layer of clay- and iron-enriched soil at 0.5–1.0 m depth that allows water to saturate the upper soil during wet winters. The differences in soil depth leads to huge differences in the forests, with tree heights dropping from 30 m on the deep soils to c. 3 m on the shallow soils. The climate is the same across terraces, but slow changes in the depth of the ecosystems (with concomitant changes in soil fertility) led to differences in vegetation and soils that are as profound as those encountered across a very large climate gradient. The critical zone (see Richter & Billings) for understanding the ecological factors that distinguish these ecosystems extends to variable depths in the soils, as well as to linkages across the landscape over very long periods of time. How deeply do most ecosystems extend into the soil? Depending on physical and chemical aspects of soils, trees roots are commonly found at 3–6 m depth, with some cases reporting over 20 m depth (Stone & Kalisz, 1991; Schenk & Jackson, 2002). A plantation of Eucalyptus trees in Brazil reached 25 m in height in 5 years, while roots grew to a depth of 18 m, and more than half the length of fine roots in these Eucalyptus forests occur below 1 m (Laclau et al., 2013). These deep roots provide substantial water uptake (especially in the dry season) and high uptake capacities for calcium and potassium (da Silva et al., 2011). Deep soils are quite important at the other end of the environmental gradient as well; one study of arid and semi-arid ecosystems found that roots deeper than 1 m in the soil dominated the uptake of calcium, potassium, and phosphorus (McCulley et al., 2004). The important dimension of depth of an ecosystem can also depend on the local vegetation type. The rooting depth of an ecosystem dominated by grasses or herbs may deepen by several meters if woody plants become established (Jackson et al., 2000). The relationship between plants and soils is reciprocal, and plants can alter soils (and subsoils) even more deeply than the extent of roots. The physical and chemical environment deep belowground shows the influence of plant-mediated changes in hydrology and chemistry. Plants strongly influence the water regime deep below the soil surface; the presence of water directly affects chemical reactions and activities of microorganisms, and indirectly influences gas permeability and the oxidation/reduction status. Biological activity within soils generates large amounts of CO2. Upper soils often have CO2 concentrations of 1%, increasing by several fold with depth below the soil surface (Richter & Markewitz, 1995). The combination of CO2 and water yields carbonic acid, a major driver of long-term weathering of minerals. Mineral weathering in turn generates alkalinity that can be transported to streams and lakes, further extending the depth of the influence of biological processes. The spatial extent of ecological processes and interactions is incomplete, and perhaps insufficient, unless the depth extends through the critical zone. How far do ecosystems extend in time? This dimension of the critical zone (see Richter & Billings) may be more problematic than the issues of ecosystem depth. Highly dynamic processes often receive more attention than very slow processes. For example, thousands of studies have examined the rate of mass loss of plant litter from mesh bags, typically finding that half of the mass is lost within one to several years (Berg & McClaugherty, 2014). Short-term decomposition may have implications for the rates of nutrient supply to microbes and plants within a single year. However, decomposition rates may provide little if any insight on the processes that influence the long-term accumulation of organic matter in soils (Schmidt et al., 2011). Most soils contain organic matter that formed over a period of hundreds or thousands of years, yet these soils may have high flow-through of carbon atoms that cycle from plant detritus to the atmosphere with an average residence time of a few years or decades. The stabilization of very small proportions of the carbon added to soils each year accounts for long-term patterns of soil carbon accumulation, and these changes are essentially invisible in short-term studies. Similarly, the long-term processes of soil weathering and horizon development in the ecological staircase in California would have been overlooked in any short-term rate of change, but were fundamental to the current ecological characteristics of the sites. Richter & Billings tackle these issues and stress the opportunities offered by integrating the classical concept of ecosystems with the Earth science concept of critical zones. The critical zone idea harks back to Tansley's advocacy for viewing ecological processes and interactions comprising ‘one physical system’ that encompasses sufficient area, depth, and time to embrace all the important dynamics of an ecosystem. Richter & Billings note that many environmental monitoring programs invest heavily in measuring shallow and rapid processes. Without an adequate conceptualization of ecosystems extending through critical zones, these programs miss the deep and slow processes that can shape long-term dynamics. Our world has largely been shaped by deep, slow, and long-lasting processes and interactions. This is a good time to dig deeper into Tansley's ecosystem concept and extend it to embrace full critical zones.