Influence of Hydroscape Configuration on Watershed Carbon Signatures
Carbon plays a central role in most ecosystem biogeochemical processes and composes approximately half of all living tissue (Schlesinger 1997). Consequently, understanding the underlying mechanism controlling watershed carbon signatures (i.e., form, amount, and temporal patterns) has been a focus of many ecological studies. More recently, there has been a growing interest in carbon cycling among scientific and political entities due to carbon dioxide's (CO2) role in global climate dynamics (e.g., Cox et al. 2000, Fung et al. 2005).
At a global scale, a majority of carbon is buried in sedimentary rocks. Dissolved inorganic carbon (DIC) in the ocean is the largest near-surface pool representing about 95% of all active pools (Schlesinger 1997). Soil carbon pools represent the next largest compartment (~3.8%), of which one-third is stored in northern peatlands (Gorham 1991). Aside from peatlands, aquatic ecosystems represent an insignificant global carbon pool and thus have largely been overlooked by scientists working on global or regional scales (ASLO 2003, but see Bridgham et al. 2006). However, these ecosystems are highly active biogeochemically and may exert disproportionately large effects on elemental mass balances and cycling rates (ASLO 2003). Consequently, understanding the effects that these ecosystems have on carbon cycling is critical for quantifying carbon dynamics at watershed and regional scales.
Landscape ecology has continually emphasized that spatial heterogeneity and pattern are important drivers of ecological processes at all scales (Chapin et al. 2002, Turner 2005) and the study of aquatic ecosystems has benefited immensely from being viewed in a landscape perspective. The river continuum concept (RCC) provides a general perspective as to how the physical and ecological characteristics of lotic ecosystems should change from headwaters to the ocean (Vannote et al. 1980). More recently, lakes have also been evaluated along a spatial gradient similar to streams. The lake landscape position concept (LLC; Kratz et al. 1997) provides a framework for understanding the mechanisms driving ecological variation among lakes as a function of lake position within a watershed.
The movement of water across a landscape links aquatic ecosystems together as cohesive networks or hydroscapes. Consequently, the biogeochemical signatures of any individual aquatic ecosystem represent not only in situ processes, but also processes occurring in upstream linked systems (Jones and Mulholland 1998). Despite the prevalence of linkages among aquatic ecosystems, very little is known regarding how configurations of aquatic elements embedded in the terrestrial landscape function to control the movement of chemicals and energy (ASLO 2003). Thus, the overarching goal of this proposed research is to address how the configuration of different aquatic ecosystems embedded in a hydroscape influences watershed carbon signatures.
Although concepts derived from landscape ecology fundamentally structure our understanding of terrestrial ecosystems, similar perspectives aquatic landscapes are less common (but see Wiens 2002). In this study, I am proposing taking a hydroscape perspective that emphasizes studying aquatic ecosystems in the context of complex, linked configurations of different aquatic elements. Hydroscape configuration is a function of the different aquatic ecosystems present and how those ecosystems are linked together. Under this framework, each aquatic ecosystem represents a distinct functional unit or patch in which physical and biological processes modify the biogeochemical signatures transmitted to downstream ecosystems. Spatial configuration of patches and interactions between patches influence ecological patterns and processes at the landscape scale (Chapin et al. 2002). Thus, I argue that watershed carbon signatures are not equal to the sum of the parts (e.g., % wetland or lake), but also reflects the spatial arrangement of these ecosystems in the broader hydroscape. Put another way, I hypothesize that hydroscape configuration alters the amount, form, and temporal patterns of carbon.
Although limited attention has been given to how the spatial configuration of aquatic ecosystems influences watershed carbon signatures, past studies can provide insight into how different systems (e.g., lakes, wetlands, and streams) influence carbon signatures. A majority of watershed carbon research has emphasized correlations between concentrations or yields of organic carbon and wetland extent (e.g., Gergel et al. 1999, Mulholland 2003). The role of wetlands as sources of carbon to downstream ecosystems is well established (Hope et al. 1994, Aitkenhead and McDowell 2000). Aside from dissolved organic carbon (DOC), wetland influence on other forms of carbon (i.e., DIC, CO2, and CH4) is less understood (Hope et al. 1994, Hope et al. 2001, Worrall and Burt 2005) and difficult to quantify due to rapid degassing (Hope et al. 2004).
At a landscape scale, compared to wetlands, far less is known regarding how lakes influence the flux of carbon to downstream ecosystems. Lakes appear to be DOC sinks and reduce downstream carbon fluxes (Dillon and Molot 1997, Canham et al. 2004, Larson et al. 2007). Results from these studies suggest that internal processing of DOC in lakes can reduce annual fluxes by 70% - 90%. On the other hand, studies of beaver ponds have suggested that they can increase DOC concentrations in streams (Margolis et al. 2001). Consequently, the influences of lentic systems on watershed carbon dynamics are somewhat ambiguous. However, processing of DOC by abiotic (e.g., photobleaching) or biotic (e.g., microbial respiration) mechanisms may require considerable time to cause significant changes (Osburn et al. 2001), and watershed features that slow the movement of water may serve as spatially abrupt transformers of carbon (Larson et al. 2007).
Stream processes also have the potential to alter carbon signatures; however the residence time of water in streams is typically substantially less than lakes or other standing bodies of water. Metabolism, gas exchange, and photobleaching can alter the amounts and forms of carbon in streams (Dawson et al. 2004, Worrall et al. 2006). Gas exchange with the atmosphere has the potential to significantly influence dissolved CO2 and CH4. In a Scottish stream, exchange with the atmosphere resulted in significant loss of CO2 and CH4 inputs from peat layers within a few hundred meters of the source (Hope et al. 2001). Peat-derived DOC is generally considered to be relatively recalcitrant and only about ca. 20% of DOC in streams is typically bioavailable (Thurman 1985). Although DOC is generally considered to be recalcitrant, Worrall et al. (2006) reported some processing of DOC over 28 kilometers in a peat-dominated stream. Another study in the Adirondacks found that streams did not process substantial quantities of DOC and the streams primarily served as DOC transporters (Canham et al. 2004). Consequently, the relatively rapid loss of CO2 and CH4 (100s of meters) relative to DOC (10,000s of meters) suggests that stream processes have the greatest effect on inorganic carbon and little influence on DOC.
The amount of wetlands in a watershed is generally the strongest predictor of DOC concentrations in both lakes and streams contained within the watershed (Mulholland 2003, Xenopoulos et al. 2003). Although wetland cover can explain a significant proportion of DOC, wetlands do not explain all of the variance and the strength of that relationship can vary substantially. In a study of 119 lakes Gergel et al. (1999) reported that wetland cover explained 26 - 33 % of DOC concentrations in northern Wisconsin lakes. In the same study, the authors observed that wetlands explained 18 - 69% of DOC concentrations in streams during the spring and fall respectively. Xenopoulus et al. (2003) reported that the proportion of wetlands explained 30% of the variation in DOC concentrations in the Upper Great Lakes Region, USA and 36% of the variation globally (n = 357). In streams, wetland cover generally explains more than 50% of the variation in DOC (Mulholland 2003), although some studies have reported correlations as high as 85% (Aitkenhead et al. 1999).
Although none of these studies focused on how different configurations of aquatic ecosystems influence watershed carbon dynamics, they provide a framework to begin thinking about how different configurations of these functional units may influence watershed carbon signatures. I suggest that some of the variation in carbon/wetland relationships is due to the watershed's hydroscape configuration. For simplicity, consider a model in which wetlands are the only source of DOC, streams transport DOC, and lakes are DOC sinks (e.g., see Canham et al. 2004). All else being equal, if wetland-derived DOC is transported into a downstream lake at the base of a watershed, fluxes of DOC from the watershed would be lower than the same watershed if the lake is situated upstream of wetland even though carbon inputs from the wetland are identical.
The importance of hydroscape configurations as a driver of biogeochemical signatures may vary in different regions. In many regions lakes and wetlands are not common features of the landscape and hydroscapes are generally thought of as stream drainage networks. As a result of the prevalence of these types of hydroscapes, some researchers have been exploring how channel networks shape riverine habitats (Ward et al. 2002, Benda et al. 2004). However, in areas with relatively low topography, hydroscapes become substantially more complex. For example, the Northern Highlands Lake District (NHLD) in north-central Wisconsin and Upper Peninsula Michigan is characterized by relatively low topography and approximately 38% of the land surface is covered by aquatic ecosystems (Peterson et al. 2003). Water-rich areas like the NHLD are also often carbon rich landscapes because of saturation of soils and development of extensive wetlands (Gorham 1991). Consequently, the hydroscape configuration of ponds, lakes, wetlands, and streams may be an important driver of watershed carbon signatures in these regions.