Linking soil moisture balance and source-responsive models to estimate diffuse and preferential components of groundwater recharge

Abstract

Abstract.Results are presented of a detailed study into the vadose zone and shallow water table hydrodynamics of a field site in Shropshire, UK. A conceptual model is presented and tested using a range of numerical models, including a modified soil moisture balance model (SMBM) for estimating groundwater recharge in the presence of both diffuse and preferential flow components. Tensiometry reveals that the loamy sand topsoil wets up via preferential flow and subsequent redistribution of moisture into the soil matrix. Recharge does not occur until near-positive pressures are achieved at the top of the sandy glaciofluvial outwash material that underlies the topsoil, about 1 m above the water table. Once this occurs, very rapid water table rises follow. This threshold behaviour is attributed to the vertical discontinuity in preferential flow pathways due to seasonal ploughing of the topsoil and to a lower permeability plough/iron pan restricting matrix flow between the topsoil and the lower outwash deposits. Although the wetting process in the topsoil is complex, a SMBM is shown to be effective in predicting the initiation of preferential flow from the base of the topsoil into the lower outwash horizon. The rapidity of the response at the water table and a water table rise during the summer period while flow gradients in the unsaturated profile were upward suggest that preferential flow is also occurring within the outwash deposits below the topsoil. A variation of the source-responsive model proposed by Nimmo (2010) is shown to reproduce the observed water table dynamics well in the lower outwash horizon when linked to a SMBM that quantifies the potential recharge from the topsoil. The results reveal new insights into preferential flow processes in cultivated soils and provide a useful and practical approach to accounting for preferential flow in studies of groundwater recharge estimation.

Avaliação de modelo de balanço hídrico com base na estimativa da recarga potencial

“…A complexidade e onerosidade do monitoramento dos movimentos da água subterrânea inviabiliza obter estimativas precisas da recarga em bacias hidrográficas. Uma alternativa válida é a modelagem, diversos tipos de modelos hidrológicos são usados para estimar a recarga (Cuthbert et al, 2013).…”

Measurement and simulation of subsurface tracer migration to tile drains in low permeability, macroporous soil

“…The longer breakthrough times in the current study may be the result of two major factors: irrigation intensity, and destruction of macropores in surficial soils. Initiation of preferential flow in macropores has been termed source responsive, in that it is related to water inputs and soil moisture conditions [Cuthbert et al, 2013]. In this study, irrigation inputs were managed in order to minimize surface ponding, which may have reduced travel times from surface to the tile drains.…”

Modified APEX model for Simulating Macropore Phosphorus Contributions to Tile Drains

“…Current widely used models that couple agricultural management to hydrology and water quality simulate either MEM or MDM pathways, or do not account for macropore flow at all (Chung et al, 1992; Youssef et al, 2005; Šimůnek et al, 2008; Blomback and Persson, 2009; Radcliffe et al, 2015). A recent review by Radcliffe et al (2015) highlights that ADAPT, HYDRUS, ICECREAMDB, and some P Indices incorporate field‐scale macropore flow to some degree, while DRAINMOD, APEX, and PLEASE do not.…”

“…A recent review by Radcliffe et al (2015) highlights that ADAPT, HYDRUS, ICECREAMDB, and some P Indices incorporate field‐scale macropore flow to some degree, while DRAINMOD, APEX, and PLEASE do not. The ADAPT model uses a simplistic description of MDM flow as a function of clay content and potential evapotranspiration demand deficiency, similar to the approach used in SWAT (Chung et al, 1992; Arnold et al, 2005). The HYDRUS and ICECREAMDB models simulate MEM flow with varying levels of complexity, from low‐parameter approaches to dual‐permeability numerical approximations of Richard's equations that require calibration of up to 16 parameters (Radcliffe et al, 2015).…”