Title: Water in the desert: Introduction to special section on River and Riparian Biogeochemistry
Abstract: [1] Perennial rivers and streams located in arid and semiarid landscapes of the southwestern United States are unique ecosystems where continual water availability results in hotspots of biogeochemical activity and biological diversity in an otherwise water-limited environment [Grimm et al., 2003; Stromberg, 2001; Stromberg et al., 1996]. These ecosystems are characterized by sharp spatial transitions from xeric uplands to narrow mesic riparian environments and aquatic ecosystems, and sharp temporal transitions with increases in discharge of several orders of magnitude during summer monsoon floods or spring snowmelt. These pronounced transitions in both space and time provide a compelling example of how hydrology defines aquatic and associated riparian ecosystems [McKnight, 1998], providing a unique environment to examine the coupling of water, carbon, and nutrient cycles. [2] The prominent role of hydrology in controlling both spatial and temporal transitions results in the potential for dramatic changes in semiarid perennial rivers and riparian systems due to small changes in the water balance as a result of both human activities and climate variability [Goodrich et al., 2000; Webb et al., 2007]. Stream discharge during the dry season is supplied by regional and alluvial groundwater which maintains low discharge, but provides a constant water source [Leenhouts et al., 2006; Goodrich et al., 1997]. In many arid and semiarid environments, groundwater pumping has lowered the water levels that support baseflow [Webb et al., 2007; Pool and Coes, 1999] resulting in a transition from perennial to ephemeral stream flow and associated changes in vegetation structure and productivity [Stromberg et al., 1996]. Elsewhere, riparian vegetation has increased due to human management for flood control, reservoir management for irrigated agriculture and discharge of wastewater although often with altered species diversity [Webb et al., 2007]. As human populations grow and the demand for limited water supplies impacts these ecosystems, there is an urgent need to quantify how hydrologic variability affects ecosystem structure and function. [3] River and riparian ecosystems in semiarid and arid environments share a number of similarities with their counterparts in more mesic environments. Structural similarities include the transitions from upland ecosystems that depend on precipitation to riparian vegetation that relies on the groundwater that supplies stream baseflow, and then from the terrestrial riparian environment to the aquatic ecosystem. Functionally, river and riparian ecosystems in both mesic and xeric environments are characterized by high levels of biological diversity and high rates of biogeochemical cycling. In all river and riparian ecosystems the advective flux of water in both the surface and subsurface underscores the open nature of biogeochemical cycling and is responsible for supplying the reactants that promote high levels of biogeochemical cycling as water moves from either the surface or groundwater through the hyporheic zone [Jones and Mulholland, 2000]. [4] There are, however, pronounced differences between xeric and mesic river and riparian ecosystems. Among these are: extreme variability in the timing and amount of primary productivity between uplands and the river/riparian system; N rather than P limited aquatic ecosystems [Grimm, 1987]; a general lack of hydrological connectivity between uplands and the river/riparian system [Fisher et al., 1998]; extreme temporal variability in discharge; and large temporal and spatial separation in the source of groundwater recharge and baseflow to the stream, with recharge typically occurring tens to hundreds of km away in surrounding mountains and taking hundreds to thousands of years to reach the river and riparian ecosystem [Eastoe et al., 2004; Plummer et al., 2004]. These differences present challenges for extrapolating results from the majority of river and riparian research, which, with a few notable exceptions like Sycamore Creek AZ [Fisher and Grimm, 1985; Grimm et al., 2003], has been performed in wetter climes, to understanding the structure and function of river and riparian ecosystems in arid and semiarid environments. [5] The purpose of this special issue is to evaluate how these differences are expressed in the biogeochemical function of river and riparian ecosystems in semiarid environments. Four of the papers focus on the San Pedro River in southeastern Arizona and were partially supported by SAHRA, the National Science Foundation funded center for Sustainability of semi-Arid Hydrology and Riparian Areas. The first paper, by Baillie et al., describes variability in sources of water to the San Pedro River, thus providing a needed hydrologic framework for the next three papers which examine how hydrologic variability affects C and N cycling. Meixner et al. [2007] presents high temporal resolution data on both water source and dissolved C and N during summer monsoon flood events, Brooks et al. [2007] examines seasonal variability in both dissolved and particulate C and N flux associated with changing water sources, and Brooks and Lemon [2007] describes spatial variability in dissolved C and N cycling associated with changes in hydrology and water source. The final paper, by Hultine et al. [2007], brings the issue back to water flux, but from the perspective of how physiological characteristics of riparian vegetation affect hydrological fluxes. Together, these papers represent a maturing, interdisciplinary understanding of riparian ecosystem structure and function in semiarid environments. [6] Baillie et al. [2007] used a set of geochemical tracers to address the following questions: “What are the contributions of different water sources (e.g. local recharge during monsoon flood events versus inflow of basin groundwater) to riparian groundwater and river baseflow?” and “How does the spatial variability in water sources relate to gaining and losing reaches along the river?” Within the San Pedro Basin, there are two dominant riparian water sources, basin groundwater which is largely mountain front recharge of winter precipitation along the boundaries of the basin, and recharge of monsoon floodwaters in the river and riparian system. Using the distinct geochemical composition of these waters, Baillie et al. [2007] demonstrate that riparian groundwater composition varies between gaining and losing reaches. Locally recharged monsoon floodwater comprises 60 to 85% of riparian groundwater in losing reaches whereas gaining reaches contain only 10% to 40% monsoon recharge. Baseflow, sampled year round, also contains a significant component of monsoon floodwater ranging from 80% on the upstream end and decreasing to 55% after passing though several gaining reaches. [7] The large spatial scale and potential magnitude of these surface water-groundwater exchanges documents a critical process in the hydrology and biogeochemistry of semiarid and arid riparian systems. Because flood water often contains elevated nutrient loads, this flux may represent a major source of nutrients to the riparian system, particularly in those losing reaches of the river where the monsoon floodwaters were found to dominate riparian groundwater. From a management perspective, the recognition of the importance of flood driven recharge highlights the challenges involved in balancing the overall basin groundwater budget. In addition to the potential impacts of groundwater pumping one must consider the possible changes in surface runoff that may result from climate change as well as changes in land use and land cover. Consistent monitoring of these geochemical tracers may serve water managers by identifying when changes in riparian water sources occur and whether these changes are due to human activity or climate variability. [8] Focusing on the summer monsoon season, Meixner et al. [2007] demonstrate that both groundwater and soil water enter the stream during flood peaks, and that this mechanism is a likely avenue for solute loading to the stream. By comparing wet and dry monsoon seasons they demonstrated that Cl could be used to conservatively trace this solute input and thereby infer biogeochemical processes in riparian soils. For example, nitrate accumulation in soil water during the relatively dry winter of 2002 resulted in floodwater and soil-water nitrate concentrations in summer 2002 that were higher than those measured during summer 2001 floods, which followed a wetter winter. Their data suggest that dry conditions both promoted the in situ production of nitrate, and inhibited the movement and/or biogeochemical cycling of this highly concentrated and potentially mobile pool of soil solutes from the soil to the stream until the onset of the summer monsoon. Analyses of later floods further suggested that solute loading, particularly nitrate, was closely related to the hydrological connectivity between soils and surface water, and to the residence time of water in soils and the subsurface. Meixner et al. [2007] also sheds light on a potential mechanism for the flood flow dominance found by Baillie et al. [2007] with flood flows displacing existing soil and groundwater and flood waters subsequently returning to the stream, potentially many months later. [9] Working at the same site, Brooks et al. [2007] performed a detailed characterization of how inter- and intra-annual variability in discharge affects the concentration and flux of both dissolved and particulate organic matter and inorganic nitrogen over an 18-month period that included two summer monsoon seasons. Although Dissolved Organic Matter (DOM) dominated carbon and nitrogen flux during the non-monsoon season, Particulate Organic Matter (POM) was found to dominate flux during the summer monsoon. Concentrations of Dissolved Organic Carbon and Nitrogen (DOC, DON), and NO3-N typically increased 2- to 10-fold while POM increased by 2 to 3 orders of magnitude during the monsoon season. Concurrent increases in concentration and discharge resulted in 96% of the annual organic carbon (DOC + POC) and 97% annual nitrogen (DIN + DON + PON) flux occurring during the monsoon season with the majority of both organic carbon (92%) and total nitrogen (83%) being transported as particulates. Over the entire study there was no predictive relationship between discharge and dissolved solute concentrations which suggests that storage and release of flood water in the alluvial system may impart a long-term biogeochemical memory on the ecosystem. In other words, solutes flushed to the stream in one year may have been stored and modified in the alluvial system for a year or more, similar to the inferences drawn by Meixner et al. [2007]. [10] A surprising aspect of this study was the preponderance of POM to the carbon and nitrogen budget of the stream. Specifically, POM transported during the monsoon season may be a major source of carbon and nitrogen to the stream ecosystem, potentially fueling heterotrophic activity in sediments and the hyporheic zone after the monsoon season and associated floods have ended. Unlike DOM however, there were predictive relationships between flow and particulate carbon and nitrogen concentrations which varied by season. In general, earlier floods transported relatively more organic matter per unit discharge, presumably due to earlier floods transporting surface litter that had accumulated on relatively dry soil since the previous monsoon season. [11] The results of Brooks and Lemon [2007] highlight the importance of spatial variability in hydrologic flow paths on constraining the biogeochemical cycling of nitrogen and carbon in this semiarid riparian system. Their synoptic surveys of the San Pedro River indicate that regional-scale variability in groundwater discharge affects nutrient concentration and cycling by limiting hydrologic connectivity and allowing nutrients to accumulate in reaches with limited groundwater input. These results are consistent with both the Baillie et al. [2007] analysis of water source and the Meixner et al. [2007] study of soil flushing of nutrients during the monsoon. Brooks and Lemon's [2007] findings further suggest that the stream and alluvial system has a memory of antecedent conditions. Reaches with limited groundwater input during the dry season tend to be N sources during the monsoon season, whereas reaches with greater hydrologic connectivity during dry seasons are nitrogen sinks. Results from both Baillie et al. [2007] and Meixner et al. [2007] suggest the mechanism for this observation is the flushing of water and nutrients originally recharged to the alluvial aquifer of losing reaches during previous floods. [12] Taken together, these first four papers suggest that the spatial and temporal scales of biogeochemical cycling in this semiarid river and riparian system are larger and longer than the typical kilometer length reach study. The primary reason for this increase in scale is the presence of an alluvial aquifer that is variably and dynamically coupled to surface water ultimately resulting in changing contributions of groundwater and surface runoff. Presumably solute loading to, processing within, and release from this alluvial aquifer impart a memory to the ecosystem where the biogeochemical response at any location reflects both current and antecedent conditions. [13] Just as the results above suggest a dynamic coupling between hydrologic flow path and biologically mediated cycling of carbon and nutrients, results from Hultine et al. [2007] clearly illustrate the impact of biological variability on hydrological fluxes, specifically evapotranspiration (ET), which is a major water flux in these semiarid environments. Using sap flow data, they demonstrate that female A.negundo trees exhibit higher sap flux densities than their male counterparts, irrespective of location or predawn water potential. These data suggest that the gender of riparian vegetation can impact ecosystem water fluxes and consequently biogeochemical cycling quite differently. These results demonstrate the importance of establishing a link between ecophysiology and population biology in order to predict outcomes of large-scale changes in vegetation structure, and its consequence on water, carbon, and nutrient cycling in response to global change scenarios. The authors note that patterns of gender specific water use in A. negundo may reflect larger scale patterns of gender effect on ET fluxes in North American riparian systems given that other dominant riparian tree species are also dioecious including Populus and Salix spp. prevalent along the San Pedro River. Therefore, characterizing species assemblages alone may not be adequate for developing fine-scale relationships between riparian ecosystem function and population structure. [14] This special issue and four of these papers were supported by SAHRA (Sustainability of semi-Arid Hydrology and Riparian Areas) under the STC Program of the National Science Foundation, agreement EAR-9876800. The mission of SAHRA is to promote sustainable management of water resources by conducting water resources-related science in the context of critical water management issues of semiarid and arid regions. One of the critical issues in semiarid regions is how to manage and preserve riparian ecosystems as both population and the demand for water grow. SAHRA's research in the structure and function of riparian ecosystems is designed to develop fundamental, process-level understanding in three areas: (1) determining the water balance of riparian systems; (2) evaluating ecosystem dynamics and values; and (3) understanding nutrient and solute sources and cycling.