A complex-systems approach to the design and evaluation of ecosystem restoration strategies

Many ecosystems are resilient against restoration efforts because of physical-biological feedbacks that maintain the degraded condition as an alternate stable state. For example, my earlier work revealed that decades of reduced flow through the Florida Everglades have led to the encroachment of sawgrass into free-flowing open water zones but that simply restoring flow will not re-open the flow pathways. Rather, the presence of vegetation in the flow-ways inhibits the erosion of sediment and the re-cutting of channels, even under enhanced velocities. Thus, restoration of open-water flow in the Everglades will require additional measures, such as reducing the abundance of sawgrass through natural (e.g., fire followed by flooding) or managed processes. In the Everglades and other environments where strong physical-biological feedbacks are present, restoration strategies need to be grounded in a firm understanding of the system’s dynamics.

Big Flat pond 1

A pond associated with pond-and-plug restoration of wet meadows, Plumas National Forest, CA

Another environment that experiences strong physical-biological feedbacks and is a target for restoration efforts and a major focus of ESDL research is wet meadows, valued for their water storage functions, water purification (e.g., through the removal of nitrate), and habitat provisioning. Stratigraphic records indicate that wet meadows were once prevalent in piedmont (mid-gradient) floodplains throughout North America and Europe. However, the loss of this landscape type has been equally widespread, and many of these landscapes have been replaced with single-channel, incised streams that are poorly connected to their floodplain. Throughout the Sierra Nevada, former wet meadows that transitioned to xeric, sagebrush-dominated landscapes following decades of grazing and deforestation have become targets for restoration. Common goals of wet meadow restoration projects include reestablishment of diverse, herbaceous wet meadow vegetation, providing habitat for fish, amphibians, and avian life, and sustaining higher baseflow for extended periods following snowmelt. However, the most common type of Sierra Nevada wet meadow restoration technique, termed “pond and plug,” transforms ecosystems into a novel state, different from the past or present condition. Sustainable restoration strategies for wet meadows will require an understanding of the suite of factors and interactions that have led to their degradation and how those factors interact with present-day landscape conditions. Given the former ubiquity of wet meadows and the similar ubiquity of their loss, achieving a general understanding of the factors sustaining and degrading wet meadows requires a geographically broad, process-based approach.

Healthy wet meadow tributary showing a distinct hyporheic zone, Plumas National Forest, CA

Healthy wet meadow tributary showing a distinct hyporheic zone, Plumas National Forest, CA

Several projects that the ESDL is engaged in focus on understanding the mechanisms for landscape development in alluvial valley bottoms so that mechanistic models can be formulated. Those numerical models can then be used to evaluate the sets of conditions under which wet meadow environments are a stable landscape configuration and test hypotheses about perturbations that cause shifts to alternate configurations such as deeply incised, single-threaded channels. Ongoing field work focuses on delineating the potential range of flow-vegetation-sediment-nutrient interactions in wet meadows and their ability to impact landscape evolution. We are conducting surveys of biological and physical variables in meadows that have recently undergone restoration that reset landscape conditions by removing topsoil and previous vegetation and reconfiguring channels. By sampling these landscapes soon after a resetting event (i.e., “restoration”), we have a higher probability of capturing the diversity of potential biophysical feedbacks and observing their influence on landscape development over the span of a few years. In our sampling, focused on wet meadows in California, Wisconsin, and Pennsylvania, we are testing the hypothesis that vegetation patches with a limited number of distinct physiologies engineer their environments in ways that are predictable given knowledge of valley slope, annual rainfall, and groundwater inflow. At early stages of landscape development, we hypothesize that these feedbacks govern channel number and depth and hence the frequency and extent of flow across the floodplain.