Spatial distribution of transit times in montane catchments: conceptualization tools for management

✍ Scribed by C. Soulsby; D. Tetzlaff; M. Hrachowitz

• Publisher: John Wiley and Sons
• Year: 2010
• Tongue: English
• Weight: 321 KB
• Volume: 24
• Category: Article
• ISSN: 0885-6087
• DOI: 10.1002/hyp.7864

✦ Synopsis

The hydrological community is showing increasing interest in using tracers to understand catchment function by characterizing transit times (Beven, 2010). This usually requires fitting a transit time distribution (TTD) to the input-output dynamics of assumed conservative tracers which allows statistics such as the mean transit time (MTT) to be estimated (McGuire and McDonnell, 2006). Although such an estimation involves various assumptions and associated uncertainties, the MTT can be used as a fundamental metric for the integrated catchment function that gives a scale-independent index that can aid catchment classification and intercomparison (McDonnell et al., 2010). Such indices can be particularly useful in many montane areas, where data are usually particularly sparse and MTTs are often relatively short and estimates are well constrained. Additionally, MTTs can often be predicted from landscape characteristics in montane areas where topography (McGuire et al., 2005), soils (Soulsby et al., 2006a) or geology (Viville et al., 2006) act as first order controls. Thus, the catchment structure can be explicitly linked to the hydrological function in a predictive manner which allows MTTs to be estimated for ungauged basins from landscape characteristics that can be derived from a geographic information system (GIS) (McGuire et al., 2005; Soulsby and Tetzlaff, 2008). In addition, MTTs have also been shown to be closely correlated with hydrometric design statistics (Soulsby et al., 2010). This enables MTT estimation for ungauged basins to be used to also estimate flow statistics, as well as potentially providing insight into catchment-scale water storage (Soulsby et al., 2009). In the Scottish Highlands (which cover about 70% of land area), a study of 20 catchments ranging from 0•5 to 30 km 2 used multiple regression to show that almost 90% of the variability in MTTs derived from long-term tracer data could be explained by four parameters (Hrachowitz et al., 2009). These parameters related to landscape controls that can be readily extracted from a GIS: (i) responsive soil cover (RSC) (i.e. soils that generate quickflow), (ii) drainage density, (iii) long-term precipitation intensity and (iv) topographic wetness index (TWI). The resulting model could predict MTTs with a relative error of 0•26 and was independently validated in 20 additional catchments in the Scottish Highlands at scales ranging from 1 to 1700 km 2 (Hrachowitz et al., 2010a). Given the relatively robust nature of the model, it offers the potential to estimate MTTs in ungauged basins in the Scottish Highlands with a reasonable degree of accuracy. This provides a new basis for disaggregating MTTs at the mesoscale and conceptualizing how landscape characteristics interact to integrate the small-scale heterogeneity which contributes to the evolution of MTTs at larger scales. In this Commentary, we suggest that this can contribute to the provision of tools for communicating process-based hydrological