Hydrologic flushing rates drive nitrogen cycling and plant invasion in a freshwater coastal wetland model

Abstract

Coastal wetlands intercept significant amounts of nitrogen (N) from watersheds, especially when surrounding land cover is dominated by agriculture and urban development. Through plant uptake, soil immobilization, and denitrification, wetlands can remove excess N from flow‐through water sources and mitigate eutrophication of connected aquatic ecosystems. Excess N can also change plant community composition in wetlands, including communities threatened by invasive species. Understanding how variable hydrology and N loading impact wetland N removal and community composition can help attain desired management outcomes, including optimizing N removal and/or preventing invasion by nonnatives. By using a dynamic, process‐based ecosystem simulation model, we are able to simulate various levels of hydrology and N loading that would otherwise be difficult to manipulate. We investigate in silico the effects of hydroperiod, hydrologic residence time, N loading, and the NH4+:NO3‐ ratio on both N removal and the invasion success of two nonnative species (Typha × glauca or Phragmites australis) in temperate freshwater coastal wetlands. We found that, when residence time increased, annual N removal increased up to 10‐fold while longer hydroperiods also increased N removal, but only when residence time was >10 d and N loading was >30 g N·m−2·yr−1. N removal efficiency also increased with increasing residence time and hydroperiod, but was less affected by N loading. However, longer hydrologic residence time increased vulnerability of wetlands to invasion by both invasive plants at low to medium N loading rates where native communities are typically more resistant to invasion. This suggests a potential trade‐off between ecosystem services related to nitrogen removal and wetland invasibility. These results help elucidate complex interactions of community composition, N loading and hydrology on N removal, helping managers to prioritize N removal when N loading is high or controlling plant invasion in more vulnerable wetlands.

Sustained‐Flux Global Warming Potential Driven by Nitrogen Inflow and Hydroperiod in a Model of Great Lakes Coastal Wetlands

“…Recent MONDRIAN versions integrate more detailed plant physiology and competition, including clonal branching and light competition (Goldberg et al., 2017; Martina et al., 2016). Nitrogen (N) cycling in MONDRIAN was also recently enhanced by adding nitrification and denitrification (Sharp et al., 2020). The model has previously been applied in Great Lakes coastal wetlands; however, the model processes are general enough that they could be used to study inland wetlands and wetlands in other regions.…”

“…Thus, flooding enhances C and N accretion in detritus while slowing the release of both C and N from detrital pools via mineralization. Previous studies using MONDRIAN provide greater detail on C and N cycling in the model, including controls on decomposition, decomposition feedbacks on N mineralization, plant growth and uptake of N, hydrology and anaerobic zonation and their effects on C and N cycling (Currie et al., 2014; Martina et al., 2016; Sharp et al., 2020).…”

“…WL (cm) represents water level, which may be above or below the soil horizon. MONDRIAN calculates WL using a trailing average water level of 5 days to account for the delay in the switch from aerobic to anaerobic conditions (Sharp et al., 2020). The effect of temperature (°C) is f T (Equation ), where WT represents water temperature, which is calculated from air temperature ( T air ) (Equation ).…”

“…Sierszen et al (2012) used isotopes to measure the water residence time in coastal wetlands and found it to range from 0.16 to 46 days across their study sites. We estimated a wide range based on the variety of coastal wetlands in the region (Sharp et al, 2020), including 1, 10, 30, and 100 days (Table S1).…”

“…Native species were randomly distributed in the modeling area and given 14 years to become well established, while invasive (Phragmites) plants were introduced at random locations in year 15. While we do not expressly address plant invasion in the present study, previous work with MONDRIAN has shown that this produces a realistic range of plant communities over a range of N inflow levels (Currie et al, 2014;Martina et al, 2016;Sharp et al, 2020).…”