Nous présentons ici une méthode simple qui permet de mesurer simultanément l'eau et les sédiments en transit à la surface du sol en milieu agricole. Des hydrogrammes de ruissellement superficiel sont comparés pour évaluer la variabilité saisonnière et spatiale du ruissellement et de l'érosion lors de trois précipitations naturelles (1,16, 1,76, 1,78 mm h-1). Les mesures ponctuelles sur le terrain sont utilisées pour caler une équation de transport des sédiments compatible avec les modèles hydrologiques distribués. Le débit solide (Qs) est exprimé par:
Qs = k Qm lp
où Q est le débit liquide et l l'intensité des précipitations. Les hydrogrammes montrent que les débits varient dans l'espace selon la distance entre deux points d'échantillonnage et, dans le temps selon l'état de la surface. Pendant la saison, la concentration moyenne des sédiments dans l'eau et le taux de transport demeurent constants. Le transport par le ruissellement sur la parcelle est faible et semble limité par la capacité du processus à maintenir les sédiments en mouvement. Ce comportement affecte le paramètre m associé au débit liquide dans l'équation de transport, qui est en dessous des valeurs théoriques proposées dans la littérature. Le transport de sédiments est également fonction de l, qui par le biais de l'impact des gouttes de pluie sur la surface détache et facilite le transport des particules.
- hydrogramme de ruissellement,
- transport de sédiments,
- milieu agricole,
Hillslope hydrographs and sediment transport on an agricultural field
Data are needed to validate a storm by storm model of sheetwash erosion, to assess spatial and temporal variations of runoff and erosion and to calibrate a sediment transport equation on agricultural fields. This remains a major problem in the development of distributed hydrological models. This paper presents a simple method to measure simultaneously water and sediment discharges on hillslopes. Hillslope hydrographs and sediment transport rates are used to investigate spatial and seasonal variations in runoff and erosion. Measurements are also used to calibrate a sheetwash sediment transport equation compatible with distributed hydrological models. Sediment discharge (Qs) is expressed by :
Qs = k Qm lp
where Q is water discharge, S is hillslope gradient and l rainfall intensity. Parameters k, m, n and p are constants for a given context.
The experimental site is located in the Eastern Townships (Québec, Canada). It is a corn field (1 000 m2) where sheetwash erosion is active. Simultaneous measurements of water and sediment discharges are collected using hydraulically efficient samplers specially designed to minimize direct rainsplash input and to prevent sediment accumulation within the receptacle. Data were collected during three natural rainfalls with low average intensities monitored in June (l -1,76 mm h-1), September (l = 1,78 mm h-1) and October (l = 1,16 mm h-1) 1987. Because rainfall intensity varies within a precipitation, each rainfall event was subdivided into distinct measurement periods of short duration (5 to 60 min) with intensities ranging from 0,12 to 8,9 mm h-1. For each precipitation two samplers were operated simultaneously over the field. One of the sampler occupied a fixed spatial location, which allows comparison between event according to variations in vegetation cover and soil compaction. The second sampler is located at a different location for each event in order to sample different spatial contexts. Overall 80 samples with a measurable amount of water and sediments were obtained.
Firstly, our results show that peaks in sediment transport rates are in most cases associated with peaks of water discharges, which occur simultaneously or just after the maxima in rainfall intensities. For the first and second rainfall events similar hydrological and erosional responses of the field were observed at the two spatial locations. For there events, the distance between the two samplers on the field M short (respectively 7 and 4 meters). For a given event, the two hydrographs have the same shape, although the downslope hydrograph is target and there is a lag (5 min) between peak discharges. These characteristics suggest a kinematic response of runoff on the field. For the third precipitation, the two samplers are located 13 meters apart. There is an important diminution of runoff and erosion downslope. The inversion of the size (volume of water) of the hydrographs is attributed to the divergence of runoff caused by the microtopography and the presence of obstacles on the surface.
Over the season there is a difference in the hydrological response of the field. In September, the vegetation cover is dense and the mean infiltration and interception rates are high. Mean water and sediment discharges are low in comparison with those measured in June and October where the vegetation is sparse or absent (no interception). In October, the compaction of the soil surface is high and the infiltration capacity is low. Despite the tact that the mean rainfall intensity is slightly lower for this event, the highest amount of water and sediment discharges are observed. Average sediment concentration in water is constant for the three precipitation events. This suggests that the amount of loose sediments on the surface ready to be transported is always sufficient. The comparison of hill slope hydrographs on the field and at different times showed that water and sediment vary : i) in space according to slope length between sampling locations, and ; ii) over the season according to vegetation cover and soil surface properties.
Secondly, parameters of the sediment transport equation are estimated for the three events. The results show that water discharge and rainfall intensity are positively related with sediment discharge, but also that the effect of hillslope gradient is negligible. This is explained by the fact that only four low gradient values are used in this study and that a wide range of discharges are measured on each gradient. The overall measurements yields the empirical equation
Qs = 0,02 Q0,97 l0,6
There is a good agreement between water discharge, rainfall intensity and sediment transport (R2 = 0,91). Over the season, variation in mean discharges are not followed by fluctuations in the rate of sediment discharge which remained constant for the three events. Sediment transport appears to be low and under the limit of transport for rainwash erosion. As a result, the rate of increase in sediment transport (m = 0,97) is under the lower limit of the theoretical range of values (from 1,4 to 2,4) proposed in literature. Sediment discharge is also influenced by the contribution of rainsplash to particle detachment and transport. This process is evaluated by the incorporation of rainfall intensity into the sediment transport equation. The value obtained from the empirical estimation of p (0,6) is not significantly different from the theoretical value (p = 0,5) proposed in the literature. The contribution of rainfall intensity to the prediction of sediment transport is low but significant.
In conclusion, results of this experiment show that spatially and temporally distributed data can be used to increase our knowledge on runoff and erosion at the scale of an agricultural field. The rote of low rainfall intensities on runoff and erosion is also important despite the presence of vegetation. These events contribute to the transfer of particles downslope and they increase the amount of loose sediments ready to be transported on the surface. Notwithstanding the fact that the net erosion on the field is negligible, sediment transport is active and predictable using a simple sediment transport equation.
- Hillslope hydrograph,
- sediment transport,
- agricultural field,