Simulation of the annual water table dynamics of a floodplain wetland, Narborough Bog, UK

Abstract

This paper describes the simulation of a succession of annual hydro-periods describing water table variations in a British floodplain wetland, Narborough Bog. This is achieved through the application of a model, developed using MODFLOW, which requires inputs of weekly precipitation, evapotranspiration and river stage. The model reproduces the observed temporal variation in the wetland water tables illustrating the range in water table conditions experienced by the wetland, their relationship to hydrometeorological conditions, and the seasonal and annual variability in the wetland water budget. The model also estimates defuse water fluxes such as seepage flows to and from the adjacent river. The model simulations demonstrate the importance of successful representation of the stratigraphy of alluvial and organic deposits within the wetland. Model performance is evaluated for 2 years: 1991 and 1992 during a period of field instrumentation. The model is then used to reproduce the wetland hydroperiod for further 6 years: 1979, 1980, 1982, 1983, 1987 and 1988 when local data for precipitation, evapotranspiration and river stage are available. The results indicate how the wetland responds to a combination of hydrological events and emphasise the importance of precipitation—both weekly totals and its annual distribution. The model simulations reveal the importance of the water storage function of the wetland and indicate the varying relationship of the wetland to the lowland river: at low flows the river represents a base level to which the wetland water table grades. Conversely, for small periods of time wetland water tables are elevated by the effects of a marginal silt–clay levee adjacent to the riverbank. Conditions of influent seepage from river to wetland were experienced only on isolated occasions. Such results demonstrate the benefits of floodplain wetlands in moderating extremes of river flow and demonstrating the need for rivers and floodplain wetlands to remain closely integrated.

Introduction

Wetlands are dynamic landforms characterised by periods of saturation that produce a recognisable wetland substrate and biota (NRC, 1995). This general definition includes distinct wetland types that differ in their water source, nutrient flux and in the temporal pattern of water flow-through. Thus, determination of the wetland water balance provides a ready way to distinguish between different types of wetland especially as a result of the association between nutrient concentration and water source (Gilvear et al., 1993, Owen, 1995). For example, Gilman (1994) describes several applications of a water budget approach to wetlands ranging from those in high latitude wetlands sustained by precipitation to lowland wetlands. Generally, water flows are concentrated in the permeable upper horizons of a wetland giving rise to the ‘shallow flow model’ describing pathways of wetland water flow although in large wetlands vertical water movements may be significant (Reeve et al., 2000). For floodplain wetlands additional fluxes may include subsurface seepage to and from a river, infiltration of floodwater (Mitsch, 1978, Bradley and Brown, 1995) and spatially variable groundwater flows. Thus along a 1 km transect of the Insh Marshes, Speyside, wetland waters comprise varying quantities of river-water, seepage from ditches dissecting the wetland, groundwater seepage towards the middle of the transect and waters derived from hill-slope runoff and precipitation at the margins (Grieve et al., 1995). These hydrological trends are evident from hydrochemical data but are also indicated by changes in vegetation communities along the transect (Willby et al., 1998). In such situations, it is difficult to identify wetland boundaries with confidence, which has limited the number of complete wetland water budgets that have been derived.

A general limitation of wetland water budget studies is that they have typically been applied at an annual scale to illustrate the balance between water inflows and outflows, while many wetlands exhibit marked seasonal variations in hydrological fluxes producing characteristic water table fluctuations. Annual aggregation of wetland water inputs and outputs may also hide substantial periods of either water surplus and inundation or drought that together determine the cumulative or aggregated impacts on a wetland (Winter, 1988). For this reason the wetland water table exerts an important influence upon vegetation communities although in some wetland studies it has not described sufficiently (Cole et al., 1997). Periods of inundation or saturation enable a distinctive wetland flora to persist by contributing to oxygen stress or the build-up of toxic materials (Wheeler, 1999). Conversely, low water tables may increase the cover of terrestrial species enhancing decomposition and mineralisation of organic deposits (Freeman et al., 1993) that if sustained may lead to progressive degradation of the wetland (Fojt and Harding, 1995). These extremes apart many natural wetlands are characterised by significant variability in water tables that are difficult to replicate in created or managed wetlands.

This paper describes the results of a study that seeks to reproduce the temporal pattern of the wetland hydroperiod for a British lowland floodplain wetland. This is achieved through the application of a previously calibrated numerical model of a floodplain wetland, Narborough Bog, the development of which was outlined in Bradley (1996) and which is summarised in Section 2.3. The model was calibrated by simulating five discrete 10-day intervals that were characterised by negligible evapotranspiration (mean daily temperature <2.5 °C and frozen surface with mean calculated PE <0.3 mm) or precipitation (1 period), precipitation but no evapotranspiration (2 periods) and evapotranspiration but no precipitation (2 periods). The earlier paper did not consider the annual variation in wetland water tables although as the model requires summary of data for precipitation, evapotranspiration and river stage, an increase in the temporal scale is feasible. Further analysis of this model provides a means to identify the dominant hydrometeorological processes determining the wetland water table. An increase in the temporal scale to include the development of annual water table dynamics also clarifies the hydrology of the wetland in a number of ways. Firstly, the model estimates water flows through diffuse boundaries; these are difficult to determine accurately in the field but are needed to quantify the wetland water budget. Secondly, the model is based upon a governing equation combining Darcy's law with the continuity equation so that the water budget can be estimated directly from model output together with an indication of the quantity of water moving to or from storage within the wetland. Thirdly, the model has considerable potential as an investigative tool to examine the effects of future or past changes in hydrological conditions on individual wetlands.

To provide a framework for the further application of the model the following questions were used to guide additional model simulations:

• does the use of time-invariant parameters to describe hydraulic variables enable the annual wetland hydroperiod to be reproduced successfully?

• can the model represent the combined effects of precipitation and evapotranspiration?

• how significant is river stage to the maintenance of saturated conditions and/or water table stability within the wetland?

• what variability in the wetland hydroperiod is suggested by annual model runs in past years?