Topographic modelling of soil moisture conditions: a comparison and verification of two models

Abstract

Topography, as captured by a digital elevation model (DEM), can be used to model soil moisture conditions because water tends to flow and accumulate in response to gradients in gravitational potential energy. A widely used topographic index, the soil wetness index (SWI), was compared with a new algorithm that produces a cartographic depth-to-water (DTW) index based on distance to surface water and slope. Both models reflect the tendency for soil to be saturated. A 1 m resolution Light Detection and Ranging (LiDAR) DEM and a 10 m conventional photogrammetric DEM were used and results were compared with field-mapped wet soil areas for a 193 ha watershed in Alberta, Canada, for verification. The DTW model was closer to field-mapped conditions. Values of K match90 (areal correspondence, smaller values indicating better performance) were 7.8% and 12.3% for the LiDAR and conventional DEM DTW models, respectively, and 88.5% and 86.7% for the SWI models. The two indices were poorly correlated spatially. Both DEMs were found to be useful for modelling soil moisture conditions using the DTW model, but the LiDAR DEM produced the better results. All major wet areas and flow connectivity were reproduced and a threshold value of 1.5 m DTW accounted for 71% of the observed wet areas. The poor performance of the SWI model is probably because of its over-dependence on flow accumulation. Incorporation of a flow accumulation algorithm that replicates the effects of dispersed flow showed some improvement in the SWI model for the conventional DEM but it still failed to replicate the full areal extent of wet areas. Local downslope topography and hydrologic conditions seemed to be more important in determining soil moisture conditions than is taken account of by the SWI. The DTW model has potential for application in distributed hydrologic modelling, precision forestry and agriculture and implementation of environmental soil management practices.

Evaluating digital terrain indices for soil wetness mapping – a Swedish case study

“…For example, Kopecky and Cizkova (2010) preferred FD8 for vegetation mapping, while Sørensen et al (2006) concluded that using D∞ improved soil wetness mapping. Murphy et al (2009Murphy et al ( , 2011 found that T WI -based wetness maps improved from D8 to D∞, and these improvements were scale dependent, while D TW maps showed better and fairly scale-independent correlations for various soil properties, including soil and vegetation type and drainage class.…”

High Resolution Site Index Prediction in Boreal Forests Using Topographic and Wet Areas Mapping Attributes

“…We used DTW rasters created from 0.5 ha, 1 ha, 2 ha, 4 ha, 6 ha, and 10 ha catchment areas (CA), which represents the size of the upslope contributing area required to initiate a stream flow. A DTW index integrates both slope and distance, which means landscape points further from open water bodies, both horizontally and vertically, have higher DTW values and drier soils [18], while FA is calculated as the upslope drainage area contributing to the point of interest and is expressed as the number contributing of cells (m 2 ).…”

“…Murphy et al [18] found that DTW could delineate patterns of soil moisture (hydric to subhygric soil moisture regimes) in Central Alberta, and suggested usage of DTW for landscapes where belowground flow patterns are driven by surface topography. In addition, detailed physical and chemical soil properties, as well as soil, vegetation, and drainage type are found to be subject to topographic controls, and DTW with a 4 ha CA effectively estimated these in the Foothills Natural Region of Alberta [19].…”

“…It relies on the assumption that topography controls water flow. The WAM process includes removing artificial depressions in the DEM, determining flow direction, and producing stream accumulation networks based on predetermined flow channel initialization [18]. The flow initiation area (catchment area) is an arbitrary threshold which represents the upslope contributing area required to initiate a stream channel.…”

Use of Landsat images for yield evaluation within a small plot

“…The following topography attributes are used most widely for explaining topography influence on yield: digital elevation models (DEM) (Iqbal et al 2005, Murphy et al 2009), relative field elevation (Serrano et al 2013), slope (Pilesjö et al 2005), curvature (Guo et al 2012), flow accumulation (Marques da Silva and Silva 2008, Kumhálová et al 2013), topography wetness index (TWI) (Schmidt andPersson 2003, Sørensen et al 2006), distance to flow lines (Marques da Silva and Silva 2006) and compound topographic index (Momm et al 2013).…”