No abstract

Stream-riparian ecosystems are landscapes composed of dynamic interacting terrestrial and aquatic patches. Patch composition and configuration affects both the form of transported materials and the amount of nutrient retention and export. We describe spatial patterns of nutrients in the surface water of an arid-land stream using surveys conducted at 3 different scales, ranging from 30 m to 10 km in extent and from 1 m to 25 m in grain. We then relate these patterns to connections with subsurface patches at channel subunit, channel unit, and reach scales. Our objectives were to compare spatial variation in nutrients across scales, to determine the causes of downstream changes in nutrient concentration in terms of intervening patches, and to investigate whether subsurface patches at different scales behaved similarly in terms of net nutrient processing.Nutrients varied spatially at all scales sampled. The highest variation was observed in nitrate-N (NO 3 -N) in the survey with the smallest grain (CV ϭ 161%) and the lowest was observed in soluble reactive P (SRP) in the same survey (CV ϭ 17%). We hypothesized that downstream changes in nutrient concentrations were caused by upwelling of high-nutrient water from the subsurface. To test this hypothesis, we identified locations of hydrologic inputs to surface water from the subsurface using geomorphic features of the stream such as gravel bar edges (channel subunit scale), riffle-run transitions (channel unit scale), and permanent groundwater sources (reach scale). As surface water passed over these locations, nutrient concentrations generally increased, particularly during late succession when subsurface patches acted as sources of NO 3 -N at all 3 scales and as sources of SRP at the channel unit and reach scales. A hierarchical approach allowed us to decompose effects of subsurface upwellings at different scales and to consider interactions between them. Processes occurring in subsurface patches influenced surface water nutrient patterns at scales from a few meters to several kilometers.

Nonlinear models of ecosystem dynamics that incorporate positive feedbacks and multiple, internally reinforced states have considerable explanatory power. However, linear models may be adequate, particularly if ecosystem behaviour is primarily controlled by external processes. In lake ecosystems, internal (mainly biotic) processes are thought to have major impacts on system behaviour, whereas in rivers, external (mainly physical) factors have traditionally been emphasized. We consider the hypothesis that models that exhibit multiple states are useful for understanding the behaviour of lake ecosystems, but not as useful for understanding stream ecosystems. Some of the best-known examples of multiple states come from lake ecosystems. We review some of these examples, and we also describe examples of multiple states in rivers. We conclude that the hypothesis is an oversimplification; the importance of physical forcing in rivers does not eliminate the possibility of internal feedbacks that create multiple states, although in rivers these feedbacks are likely to include physical as well as biotic processes. Nonlinear behaviour in aquatic ecosystems may be more common than current theory indicates.

No abstract

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.Abstract. Nutrient availability in ecosystems is patchy both in space and in time. Whereas temporal trends have often been studied, less information exists on spatial patterns of nutrient availability, particularly in aquatic ecosystems. The goals of this study were (1) to describe and quantify patterns of nutrient concentration in surface waters of an arid land stream and (2) to compare spatial patterns of nutrient availability across nutrients and over a successional sequence.We describe changes in the spatial pattern of stream water nutrient concentrations over successional time (between floods) using quantitative measures of heterogeneity. Samples were collected from the middle of the channel every 25 m over a 10-km section of a Sonoran Desert stream during three periods: early succession (2 wk post-flood), middle succession (2 mo post-flood), and late succession (9 mo post-flood). Nutrient concentrations were extremely variable in space (coefficients of variations as high as 145%). Coefficients of variation increased over successional time and were consistently greater for nitrate-nitrogen than for soluble reactive phosphorus. Semi-variogram analysis showed that nutrient concentrations were spatially dependent on all dates, but to different degrees and over different distances. The distance over which nutrient concentrations were spatially dependent, as measured by the semi-variogram range, tended to decrease with successional time. The strength of spatial dependence, as measured by the slope of the ascending limb of the semivariogram, increased with successional time. The limiting nutrient, nitrogen, was consistently more spatially heterogeneous than phosphorus or conductivity, both in terms of patch size (range) and strength of spatial dependence.In streams, downstream transport combined with nutrient transformation produces patches of similar nutrient concentrations that are elongated compared with nutrient patches in terrestrial soils. Variation in nutrient concentration is likely to affect the spatial distribution of organisms and rates of ecosystem processes.