Topography is a major factor controlling both hydrological and soil processes at the landscape scale. While this is well-known qualitatively, quantifying relationships between topography and spatial variations of hydrologically relevant variables and other landscape characteristics such as the variation of soil properties still remains a challenging research topic. Several topographic indices, which can be computed from digital elevation models (DEMs), have been suggested to quantitatively describe terrain characteristics. In the following a review of several recent studies is given.
While there are many different topographic indices, it is also important to recognize that many of these indices can be computed in different ways. This is especially true for the topographic wetness index (TWI), also called TOPMODEL index (Beven and Kirkby, 1979). The different methods to calculate both upslope area and slope result in significant differences of the computed TWI maps and the correlations between TWI and observed hydrological and other variables (Sørensen et al., 2006). The upslope area can be computed in various ways which differ in the way accumulated area in one cell is portioned between its neighbouring cells. Seibert and McGlynn (2007) recently suggested a new method which combines the advantages of the multiple-directions approach (Quinn et al., 1991) and the D∞ approach (Tarboton, 1997). A new slope term, the downslope index tanαd was introduced by Hjerdt et al. (2004). The idea behind this index was that the local slope tanβ might not always be a good representation of the groundwater table hydraulic gradient because downslope topography more than one cell distant is not considered. In contrast to tanβ, which only considers the cell of interest and its neighbours, tanαd is defined as the slope to the closest point that is d meters below the cell of interest. Both slope estimates give similar results for small values of d, but the results differ for larger values of d (Hjerdt et al. 2004).
The topographic influence on soil properties is apparent in the soil catena concept. Seibert et al. (2007) used measured soil properties from the Swedish National Forest Soil Inventory, which is a long-term inventory of permanent sample plots from the Swedish National Forest Inventory. It includes a description of soil types and soil horizons as well as sampling of organic and mineral soil horizons for subsequent chemical analyses. The study focused on Podzols and Histosols, which provided 4 000 sample plots distributed over almost all of Sweden. Plot locations were determined accurately by GPS, which allowed the overlaying of plot data and the DEM. Topographic indices such as the topographic wetness index, TWI (ln(a/tanß)), were computed from gridded digital elevation data for all sample plots. Several significant correlations between topographic indices and soil properties could be found. The thickness of the organic layer increased with TWI and the thickness of the leached E-horizon increased with upslope area. Soil pH in the organic layer increased with TWI, while the C-N ratio decreased. Soil pH in the organic layer was also found to be higher for south facing slopes than for north facing slopes. The ratio between the divalent base cation (Ca and Mg) and the monovalent base cation (K and Na) concentrations in the O-horizon increased with TWI. These correlations confirmed the importance of topography on soil properties, although there was considerable scatter, which could be attributed to heterogeneity in the large data set.
Similar correlations were found in data from two study areas where soil samples were taken at plots randomly distributed over 25 km2 (Zinko et al., 2005; 2006). These studies also demonstrated a positive correlation between topographic wetness index and plant species biodiversity (evaluated by the number of species found in 200 m2 plots).
Stream networks derived from DEMs can also provide valuable information on landscape organisation (Seibert and McGlynn, 2005). One important observation is that most catchment area is enters the stream network in rather small headwater catchments. This has implications for catchment management as riparian buffer zones might be more important along these small streams than along larger streams. McGlynn and Seibert (2003) examined the variability in, and controls on, hillslope inputs to stream networks and the potential for riparian zones to regulate hillslope inputs and thereby both quantitatively and qualitatively buffer, or modify, stream responses to hillslope hydrology. They found that the ratio of riparian zone storage to hillslope inputs was the most important plot-scale measure of the buffering capacity of the riparian zone. One particularly important finding was that the catchment-wide proportion of the riparian area might be misleading. At the 280 ha Maimai research area that ratio was 0.14. When this ‘buffer capacity’ was calculated for each 20 m stream reach along the stream network, the values were below 0.14 for 75% of the stream length and the median was 0.06. Using the catchment-wide ratio would thus significantly overestimate the ‘effective’ riparian-to-hillslope-area ratio.
McGlynn et al (2003) and McGuire et al. (2005) both demonstrated that topography and landscape structure is a major control on catchment transit times for water. McGlynn et al. (2003) found the median subcatchment size to be correlated to mean water transit times for the Maimai catchment in New Zealand. These results were confirmed by new results from the Krycklan catchment in Northern Sweden where winter baseflow 18O was strongly correlated to median subcatchment sizes. The median subcatchmnet size is computed as the median of the accumulated catchment areas in all upstream stream cells. Typically the median subcatchment area first increases along the stream network (i.e., increases with catchment area) but than reaches some level where it remains constant even with largely increased catchment area. The median subcatchment area is a measure of stream network organisation, values are smaller for tree-like network patterns and larger when the network is less dissected. McGuire et al. (2005) used methods developed by Seibert and McGlynn (2003) and found catchment-wide median flowpath lengths and gradients towards the stream network to be best correlated with residence time.
The TWI pattern is constant in time because the index is a static representation of the landscape. Grabs et al. (2007) examined the predictions of saturated areas using this static topographic wetness index and compared the spatial predictions with temporally aggregated simulations of a distributed hydrological model. The model was calibrated against discharge measured in two small subcatchment of the Krycklan catchment in northern Sweden. After calibration the model was applied to the entire larger 68 km2 catchment. The dynamic groundwater level simulations of this model were temporally averaged and provided, thus, alternative indices of wetness distribution (dynamic indices). These indices were compared to the static topographic wetness index (TWI). Grabs et al. (2007) used the ability to spatially predict the occurrence of wetlands as a validation of the static and dynamic indices. In the lower sedimentary part of the catchment both approaches overestimate the wetness due to deeper and well drained soils. These areas were, thus, excluded from the analyses. For the moraine part of the catchment first results indicate that the dynamic approach is superior to the static TWI. One explanation was that the dynamic models allows considering temporally varying hydraulic gradients, which do not equal the surface gradients. While the use of dynamic models might be restricted to smaller areas (compared to the computation of static indices such as TWI), the results of such models might help to guide how to improve the calculation of topographic indices.
Above different examples on how to use digital elevation data for hydrological landscape analysis are discussed. While obviously other variables than topography are important for the spatial variation of hydrological processes, elevation data is usually the information which is most accessible. High-resolution elevation data, which becomes more and more available, provides additional opportunities for hydrological landscape analysis.
Beven, K. J., and Kirkby, M. J., 1979. A physically based, variable contributing area model of basin hydrology. Hydrological Sciences Journal, 24, 43-69.
Grabs, T.; Seibert, J.; Laudon, H. (2007). Modelling spatial patterns of saturated areas: a comparison of the topographic wetness index and a distributed model (oral presentation), Geophysical Research Abstracts, Vol. 9, 00894, 2007, SRef-ID: 1607-7962/gra/EGU2007-A-00894, European Geosciences Union, General Assembly 2007, Vienna, Austria, 15 – 20 April 2007.
Quinn, P.F., Beven, K.J. , Chevallier, P. and Planchon, O., 1991. The prediction of hillslope flowpaths for distributed modelling using digital terrain models, Hydrological Processes, 5, 59-80.
McGlynn, B.L., J.McDonnell, M.Stewart and J. Seibert, 2003, On the relationships between catchment scale and streamwater mean residence time, Hydrological Processes, 17: 175-181
McGlynn, B.L. and J. Seibert, 2003. Distributed assessment of contributing area and riparian buffering along stream networks, Water Resources Research 39(4), 1082, doi:10.1029/2002WR001521
Hjerdt, K. N., J. J. McDonnell, J. Seibert, and A. Rodhe, 2004. A new topographic index to quantify downslope controls on local drainage, Water Resour. Res., 40, W05602, doi:10.1029/2004WR003130.
McGuire, K.J., J.J. McDonnell, M. Weiler, C. Kendall, B.L. McGlynn, J.M. Welker, J. Seibert, 2005. The role of topography on catchment-scale water residence time, Water Resour. Res., Res., 41, W05002, doi:10.1029/2004WR003657.
Zinko, U., J. Seibert, M. Dynesius, C. Nilsson, 2005. Plant species numbers predicted by a topography based groundwater-flow index, Ecosystems, 8( 4): 430 – 441.
Seibert, J. and B.McGlynn, 2005. Landscape element contributions to storm runoff. In: Anderson, M. G. and J. McDonnell (Eds.). Encyclopaedia of Hydrologcial Sciences, John Wiley & Sons, Ltd, Chichester, 1751- 1761.
Sørensen, R. , Zinko, U. and Seibert, J., 2006. On the calculation of the topographic wetness index: evaluation of different methods based on field observations, Hydrology and Earth System Sciences, 10, 101–112.
Zinko, U., M. Dynesius, C. Nilsson, J. Seibert, 2006. The role of soil pH in linking groundwater flow and plant species richness in boreal forest landscapes, Ecography, 29 (4): 515-524.
Seibert, J. and B. McGlynn, 2007. A new triangular multiple flow direction algorithm for computing upslope areas from gridded digital elevation models, Water Resour. Res., 43, W04501, doi:10.1029/2006WR005128
Seibert, J., Stendahl, J. and Sørensen, R., 2007. Topographical influences on soil properties in boreal forests, Geoderma, in press
Tarboton,.D.G., 1997. A new method for the determination of flow directions and upslope areas in grid digital elevation models. Water Resources Research, 33(2): 309-319.