Elsevier

Environmental Modelling & Software

Volume 18, Issues 8–9, October–November 2003, Pages 839-850
Environmental Modelling & Software

Terrace erosion and sediment transport model: a new tool for soil conservation planning in bench-terraced steeplands

https://doi.org/10.1016/S1364-8152(03)00084-7Get rights and content

Abstract

Despite widespread bench-terracing soil erosion remains a major problem in Java’s uplands. To elucidate the causes for this lack of impact, runoff and erosion processes were studied at a variety of spatial scales within a volcanic catchment in West Java. Research indicated that soil loss occurs via rain splash and wash of rainfall-detached sediment by shallow overland flow from the terrace riser and bed, and via runoff entrainment of sediment deposited in the central terrace toe drain. The terrace erosion and sediment transport mode (TEST) was developed to physically describe these processes, as a function of vegetation and soil surface cover where appropriate, yet use as few parameters as possible. Runoff generation was described by the spatially variable infiltration model (SVIM) and a two parameter rainfall depth-intensity distribution was assumed to derive a simple analytical expression for storm runoff depth. In a similar manner expressions were derived for effective rainfall kinetic energy to predict rainfall-driven transport using a newly developed model, and for effective runoff rate to predict flow-driven transport using GUEST model theory. The model and its components were tested using measured runoff and soil loss from 31 sections of terrace beds or risers and from six terrace units during two seasons. The model generally performed satisfactorily and provides a useful new tool for assessing the impacts of soil conservation measures on bench terrace runoff and soil loss.

Introduction

Soil erosion constitutes a major problem in the volcanic uplands of Java, affecting the livelihood of millions of farmers whereas the sediment causes additional problems downstream. In response, past development projects have advocated the construction of back-sloping bench terraces, but despite widespread adoption the problems remain (Purwanto, 1999). Some have suggested that sources other than the agricultural hillsides continue to provide the rivers with sediment, pointing at landslides, bank erosion, expansion of paddy rice fields and contributions from roads and village areas as potential culprits (Diemont et al., 1991). However, it has also been shown that bench terracing has not always been instrumental in reducing soil loss to acceptable levels. For example, recent research in the small (125 ha) volcanic Cikumutuk catchment in West Java suggested that about two-third of the annual stream sediment load of more than 40 t ha−1 was generated on the terraced rain-fed hillsides covering three-quarters of the catchment (Purwanto, 1999, Van Dijk, 2002).

Additional research in the same area was focused on identifying and quantifying the processes leading to runoff and soil loss from the back-sloping bench terraces. Important causes for the lack of effectiveness of terracing as a means of soil conservation were shown to be related to the sparse vegetation and soil surface cover, particularly on the terrace risers, the considerable amounts of runoff generated, and sometimes the steep gradient and poor management of the terrace toe drain (Purwanto, 1999, Van Dijk, 2002). This information and additional knowledge obtained in field and laboratory experiments were used to develop a model that can be used as a tool for soil conservation planning.

The traditional ‘black-box’ approach reflected in the (revised) universal soil loss equation, (R)USLE (Renard et al., 1997) does not allow an explicit representation of the recognised processes. In particular, it fails to describe the ‘lateral’ flow (through the central drain, parallel to the slope) of runoff and sediment involved and this leads to considerable uncertainty in the use of the equation (Agus et al., 1998, Critchley, 2000). Additional problems affecting the (R)USLE in the bench-terraced Javanese environment can be ascribed to the obstruction of runoff created by Wischmeier-type plots (Bruijnzeel and Critchley, 1996) and the extrapolation of an approach designed for a temperate environment to a tropical (rainfall) climate and contrasting soils, field sizes and agricultural practices (Van Dijk, 2002).

For this reason, a different approach was taken, aimed at developing a model in which the important hydrological and erosion processes are described explicitly in a physically meaningful manner. This paper presents a brief overview of the resulting terrace erosion and sediment transport (TEST) model and its application. The experiments forming the basis of the model, the model itself and the measurements used for testing are discussed elsewhere (Van Dijk, 2002, Van Dijk et al., 2002b).

Section snippets

Study area

All research was conducted within the 125 ha head water catchment of the Cikumutuk river, situated about 40 km east of Bandung, West Java, in the middle reaches of the Cimanuk basin at an altitude of 560–740 m a.s.l. (7°03′S, 108°04′W; Fig. 1). Slopes are generally fairly steep at about 15°. The Quaternary volcanic tuffs have weathered to kaolinitic Oxisols consisting of several decimetres of permeable, well-aggregated soil on top of a less permeable, massive subsoil. About two-thirds of the

Experimental process knowledge

From October 1994 onwards various methods were used to elucidate the processes dominating erosion on the different components of the terrace, including: (i) a combination of Gerlach-type trough and ‘splash boxes’ to study detachment and transport by rain splash and shallow overland flow (wash) on the terrace risers; (ii) differently sized splash cups to study rain splash on the terrace beds; and (iii) 0.3×0.6 m2 experimental soil trays (Wan et al., 1996) exposed to natural rainfall to study

General modelling approach

Model equations were formulated with the aim to use as few parameters as possible, yet preserve a physical process description. To describe runoff generation and runoff entrainment the theory used in the GUEST model (Rose, 1993, Soil Technology, 1995) was used. Theory describing the rainfall-driven processes of rain splash and wash was newly developed. Furthermore, to enhance the models potential for further use and to restrain calculation efforts, an idealised exponential rainfall

Model application

Measurements of rainfall intensity and event runoff depth and sediment yield from 15 terrace riser, 14 terrace bed and six terrace unit plots during two consecutive wet seasons (1998–1999 and 1999–2000) were used to calibrate and test (components of) the TEST model. To do so, first depth-averaged rainfall intensity (R̆) was calculated from tipping bucket rainfall data using Eq. (1). The resulting time series of event P and R̆ was used in combination with , to model event runoff depths. Because

Runoff

Performance of the runoff model component was generally good. For terrace riser and bed plots it resulted in model efficiencies of 0.65–0.87 with an average of 0.74 (leaving aside one plot that experienced technical problems; Van Dijk, 2002). Event runoff depths were modelled with an accuracy of a few millimetres for most events. Lesser performances could be explained by the influence of the preceding dry season on infiltration characteristics at the initial stages of the rainy season.

Conclusions

The TEST model predicted runoff and soil loss from back-sloping bench terraces with encouraging accuracy. At the same time, the number of calibrated parameters remained limited, while preserving the physical consistency of the model theory. Testing of the model in a more complex environment and on different soils presents a future challenge. The explicit physical description of the various processes responsible for soil loss from the terraces provides obvious advantages over the more empirical

Acknowledgements

This work was performed within the framework of the Cikumutuk Hydrology and Erosion Research Project (CHERP) in Malangbong, West Java. Albert van Dijk was supported by a grant from the Netherlands Foundation for the Advancement of Tropical Research (WOTRO, grant no. W76-193), which is gratefully acknowledged. Further details about the project as well as cited publications by the authors can be found at http://www.geo.vu.nl/~trendy.

References (21)

There are more references available in the full text version of this article.

Cited by (41)

  • Legumes protect the soil erosion and ecosystem services

    2022, Advances in Legumes for Sustainable Intensification
  • Remote sensing for the analysis of anthropogenic geomorphology: Potential responses to sediment dynamics in the agricultural landscapes

    2020, Developments in Earth Surface Processes
    Citation Excerpt :

    Erosion processes in terraced landscapes are mostly superficial. They mainly occur by shallow overland flow from the terrace riser and bed (van Dijk and Bruijnzeel, 2003; Romero Díaz et al., 2007; Lesschen et al., 2008; García-Ruiz et al., 2013; Tarolli et al., 2014). Farmers could use the methods proposed here to identify drainage systems best suited to mitigate localized erosion and to improve water usage for irrigation (when superficial).

  • Response of soil detachment rate by raindrop-affected sediment-laden sheet flow to sediment load and hydraulic parameters within a detachment-limited sheet erosion system on steep slopes on Loess Plateau, China

    2019, Soil and Tillage Research
    Citation Excerpt :

    Kinnell (2000, 2001, 2006) also identified four detachment and transport systems operating in sheet erosion; these four are raindrop detachment and splash transport, raindrop detachment and raindrop-induced flow transport, raindrop detachment and sheet flow transport and sheet flow detachment and sheet flow transport. However, most researchers have suggested that raindrop detachment (i.e. soil detachment is caused by raindrop impact), splash transport and sheet flow transport are the major processes for sheet erosion (Wan et al., 1996; Sutherland et al., 1996; Van Dijk and Bruijnzeel, 2003; Kinnell, 2006; Fu et al., 2011; Defersha et al., 2011; Zhang and Wang, 2017). The third process (i.e. detachment by runoff) reported by Ellison (1944, 1947a,b,c) and the fourth process (i.e. sheet flow detachment and sheet flow transport) reported by Kinnell (2000, 2001, 2006) cannot appear in a sheet erosion system.

View all citing articles on Scopus
View full text