Complex networks of functional connectivity in a wetland reconnected to its floodplain. Issue 7 (28th July 2017)

• Record Type: Journal Article
• Authors: Larsen, Laurel G.; Newman, Susan; Saunders, Colin; Harvey, Judson W.
• Is Part Of: Water resources research. Volume 53:Issue 7(2017)
• Page Start: 6089
• Page End: 6108
• Publication Date: 2017-07-28
• Subjects: complex networks -- connectivity -- biotic filter -- flow restoration -- Everglades; Hydrology -- Periodicals; 333.91
• DOI: 10.1002/2017WR020375
• Languages: English
• ISSNs: 0043-1397

Abstract

Abstract: Disturbances such as fire or flood, in addition to changing the local magnitude of ecological, hydrological, or biogeochemical processes, can also change their functional connectivity—how those processes interact in space. Complex networks offer promise for quantifying functional connectivity in watersheds. The approach resolves connections between nodes in space based on statistical similarities in perturbation signals (derived from solute time series) and is sensitive to a wider range of timescales than traditional mass‐balance modeling. We use this approach to test hypotheses about how fire and flood impact ecological and biogeochemical dynamics in a wetland (Everglades, FL, USA) that was reconnected to its floodplain. Reintroduction of flow pulses after decades of separation by levees fundamentally reconfigured functional connectivity networks. The most pronounced expansion was that of the calcium network, which reflects periphyton dynamics and may represent an indirect influence of elevated nutrients, despite the comparatively smaller observed expansion of phosphorus networks. With respect to several solutes, periphyton acted as a "biotic filter, " shifting perturbations in water‐quality signals to different timescales through slow but persistent transformations of the biotic community. The complex‐networks approach also revealed portions of the landscape that operate in fundamentally different regimes with respect to dissolved oxygen, separated by a threshold in flow velocity of 1.2 cm/s, and suggested that complete removal of canals may be needed to restore connectivity with respect to biogeochemical processes. Fire reconfigured functional connectivity networks in a manner that reflected localized burn severity, but had a larger effect on the magnitude of solute concentrations. Plain Language Summary: A big part of the plan to restore the Everglades involves the removal of levees and canals that are currently serving as barriers to flow. However, doing so at a large scale prompts some concerns and questions: (1) Will phosphorus, a limiting nutrient that has the capability to vastly change vegetation communities, be mobilized and transported to downstream ecosystems? And (2) Do the barriers to flow need to be completely removed (an expensive proposition), or will partial removal suffice? To address these uncertainties, a team of scientists and engineers conducted a multi‐year experiment involving controlled flow releases into a 2 km x 2 km part of the wetland that was formerly isolated from flow. In this paper the authors use a new functional connectivity network approach to visualize how those flow releases change the way constituents in the flood waters move through the downstream ecosystem. They found that flow reconnection causes a shift in how calcium is processed by microorganisms near the flow inlet, which likely reflects a transition to species that thrive under slight phosphorus enrichment. They also found that only across portions of the landscape where barriers to flow have been completely removed does the landscape regain functional connectivity. Key Points: Functional connectivity networks delineate spatial configuration of biogeochemical processes more sensitively than traditional modeling Expansion of solute networks sensitive to biotic activity is consistent with low‐level nutrient enrichment following flow reconnection Flow restoration reestablishes spatial coherence of markers of microbial metabolism, even across former barriers to flow