SEDIMENT YIELD EVALUATION IN THE SEMI ARID REGION OF BRAZIL

Vajapeyam S.Srinivasan

Professor of Civil Engineering, Centre for Sciences and Technology

Federal University of Paraiba, Campus II, Campina Grande, 58109-970, Pb, Brazil

Tel.#55 83 310 1289, Fax. #55 83 310 1388/1011,

E-mail: srinivas@rechid.ufpb.br.

Abstract: The management of sedimentation problems in rivers and reservoirs depend greatly on the sediment yield from the basin land surface.In order to evaluate the effects of land use, slope and vegetal cover on the sediment production rates, several experimental erosion plots of 100 m²  and four micro-basins of about one half of a hectare were installed in an experimental basin located in the typically semiarid region of the State of Paraiba in the northeast of Brazil. The experimental basin known as the experimental Basin of Sume, is situated within the Representative Basin of Sume. The erosion plots and micro-basins were adequately  equipped to determine the total runoff volume as well as the sediment yield. In all the cases,the sediment yield was calculated from the mean concentration of sediments in the runoff obtained from various samples. Measurements of runoff and erosion under diverse conditions of management of the erosion plots and the micro-basins have been carried out since 1982.The paper presents an analysis of the relative influence of the various factors on surface runoff and land erosion.The most significant result seems to be the enormous degree of protection offered by the fragile native vegetation against surface erosion.This ability seems possible of restoration by letting the vegetation grow back even after complete clearance. 

Keywords: sediment yield, erosion, semiarid region

1  INTRODUCTION

The north-eastern region of Brazil is predominantly semiarid with the annual precipitation being 800mm or less. In this region, agriculture is the principal vocation of subsistence and the results depend much on the vagaries of the rainfall The techniques of cultivation are mostly manual and no special measures are taken by the rural farmers to protect the land from erosion and consequent loss of productivity. The highly variable precipitation regime in the region results in considerable land erosion and in order to compensate for any loss of production, more and more areas of virgin land covered by native vegetation are cleared and brought under cultivation. The degree of protection offered by the native vegetation against erosion is not well known and hence the harmful effects of the land clearing have not been adequately documented.

The knowledge of the amount of surface runoff generated and the associated soil loss is extremely important for any rational planning of hydrological basins. These aspects assume a significance of vital importance in the case of semiarid regions.In the northeastern region of Brazil faces two conflicting situations: one in which there is a need to maximize the surface runoff and store this water for use during the dry period, and the other in which the relatively thin soil cover needs to be protected against erosion and loss of soil nutrients. The solution to this, evidently, is proper sediment management that includes the identification of areas within the river basins that are more suitable for either runoff generation or soil conservation.

In order to evaluate the influence of the human activities as well as various natural factors such as slope and vegetal cover over the processes of runoff generation and soil erosion, an experimental basin was installed in a typically semiarid region in the state of Paraíba. The basin is located in the municipality of Sumé where the mean annual precipitation is 590 mm with a coefficient of variation of about 0.5. The annual class A tank evaporation is about 2900 mm The rainfall in the region is highly irregular and is concentrated in about three months of the year between February and May (Cadier et. al, 1983). The soil cover is quite thin, underlain by bedrock. The most predominant type of soil is Brown non-calcic vertic soil with gravel and stones. The permeability is only moderate with the maximum capacity in the range of 25 to 35 mm /h. The natural vegetation is mainly bush and small sized trees. Agricultural activities are carried out mostly in the rainy season and the natural vegetation is cleared bare in preparing the land.

2  FIELD INSTALLATIONS

The field studies were carried out at three different scales: micro-basins with an area of around 0.5 ha; standard Wicshmeier type erosion plots of 100 m² and, sample plots of 1 m². The sample plots were studied under simulated rainfall only (Molinier et al., 1987), and the others with natural rainfall events only. In all, four micro-basins and nine erosion plots of 100 m² were installed and equipped for the determination of the total runoff and total sediment yield for each of the events of precipitation recorded. Seven plots of 1m² located close to the 100 m² plots were submitted to a total of 55 runs in a rainfall simulator. The field installations were carried out between 1982 and 1986.

                                   Table 1  Physical characteristics of installed micro-basins and 100 m² erosion plots

Micro-basins

Erosion Plots

Table 1 provides the characteristics of the four micro-basins and the nine 100 m² erosion plots. . All the micro-basins were equipped with sediment and runoff collectors of 2300 l capacity terminating with a 90^° triangular weir designed to handle the maximum expected discharge of 270 l/s. Water level recorders were used to register the level of water in the collectors and the head over the weir. The plots were 22.1 m long and 4.5 m wide and had among them a variety of surface covers and slopes ranging from 3.8 to 9.3%. The runoff and the eroded sediments were directed into a 1000 l capacity asbestos cement collection tank with a calibrated bucket inside to collect the small flows. For large runoff when the tank might get full, the overflow was led into a second 1000 l capacity tank that would accumulate only a ninth fraction of the outflow from the first and the remainder was spilled over. Thus for any event, the total runoff was obtained by adding to the full capacity of the first tank, nine times the volume collected in the second. All the tanks were pre-calibrated.

3  COLLECTION OF DATA

The main interest was the volume of the total runoff and the total weight of the sediments, carried off at the outlet, from the micro-basins and the erosion plots, for each of the rainfall events. The runoff from the 100 m² plots was obtained by the volume held in the calibrated bucket in the first tank if the bucket didn’t spill over, especially, for low outflows. When the bucket overflowed but the first tank didn’t, the volume was obtained by adding the capacity of the bucket and the volume spilled into the tank. For the case in which the first tank overflowed, the total volume was obtained by adding nine times the volume held in the second tank to the capacity of the first. The quantity of sediments produced in each event was determined indirectly by sampling the sediment-water mixture and determining the concentration by weight for each of the representative volumes associated with the sample.

In the case of the micro-basins, the total outflow was determined by the volume retained in the collector tank when there was no flow over the weir. For the events in which there was discharge over the weir, the total volume of runoff was obtained by adding the collector tank volume to the outflow-hydrograph volume. The hydrographs were generated from the water level recorder charts for the events and the volume was obtained by planimetering the areas of these hydrographs. The sediment yield was obtained by adding the amount of sediments retained in the collector (obtained by means of taking various samples of the mixture) to the quantity of sediments carried over the weir in suspension. The mean concentration of the sediments in the flow passing through the weir was obtained by sampling the accumulated mixture siphoned into t auxiliary cans. Whenever possible, additional samples were collected directly from the outflow of the weirs in order to obtain a better estimation of the average concentration of the sediments.

In general, the same surface conditions were maintained in the plots 1and 4 as that of the micro-basins 3 and 4. Removal of the surface vegetation growth since the last clearance was carried out in all of them at the same time. Similarly, the mulched plots 2 and 3 were subject to identical maintenance operations.

4  RESULTS AND DISCUSSION

 The collection of data  started in 1982 and since then, more than 300 events of natural precipitation, that produced runoff in at least one of the units of the experimental basins have been registered. However, the number of events with very low runoff volume and erosion are far more numerous than those with medium to large rates. This bulk of data was  utilised to evaluate the influence of factors like the surface vegetal cover, slope and cultivation practices. While the data collected from the micro-basins served essentially to evaluate the effect of clearing the native vegetation and modelling the runoff and erosion processes (Srinivasan & Galvão 1995), the data from the 100 m² erosion plots served to evaluate the effect of slope and cultivation practices. The studies with 1 m² plots utilising the simulated rainfall was used to verify the scale effect and  the variation of infiltration rates in the experimental basin (Molinier et al, 1987).

5  INFLUENCE OF THE NATURAL VEGETAL COVER

Micro-basins 1 and 2 were maintained all the time with undisturbed  natural vegetation, while the other two  micro-basins were cleared of all vegetation and kept bare. Any resurgent vegetation was cleared periodically, as was the case with the plots 1 and 4. Thus a comparison of the data from these would give a fair idea of the influence of the native vegetation on runoff and soil erosion. In the case of the micro-basins, the difference in the runoff and erosion rates between natural and cleared ones was very large. While the basins with natural vegetation produced almost no runoff for precipitation events of up to 30 mm, the bare micro-basins produced runoff with precipitation as low as about 4 mm. In the region of study, most of the precipitation events last less than an hour and rains in successive days are uncommon even during the rainy season. The amount of runoff generated in the case of cleared micro-basins varied a lot with the initial soil moisture condition, but no such influence was noticeable with vegetated basins where only very large precipitation events produced any runoff. Figs. 1 and 2 show the observed rainfall runoff relationship for micro-basins no.1 and 4. The large scatter in Fig. 2 was found to be due to the initial soil moisture condition and by introducing a parameter representing the antecedent precipitation, it was found to be possible to define a set of such relationships (Srinivasan et al, 1988). In terms of sediment yield, the protective influence of the natural vegetation became even more significant. While for precipitations of even 30 mm and higher, the micro-basins 1 and 2 would hardly produce any erosion, the sediment yield in the case of bare micro-basins 3 and 4 was very large. For a precipitation of 34.5 mm that occurred on March 25, 1989, the run off in the four micro-basins were: 0.08, 0.014, 14.85 and 15.54 mm respectively. The corresponding sediment yields were: 0.0186, 0.334, 7483.1 and 7839.2 kg/ha. Considering the sparse and dry nature of the vegetation typical of the semiarid region, the observed differences are striking. The trend was very similar in the case of 100 m² erosion plots. The bare plots 1 and 4 always produced much higher runoff and erosion rates compared with plot 5. Further, in the case of plot 9, which was installed in an area in which the cleared native vegetation had been allowed to grow back during at least 5 years, the runoff and erosion rates were only slightly higher than in plot 5, thus confirming the enormous protective influence of the natural vegetation of the region.

Mulching also reduced, significantly, the runoff and erosion rates. The plots 2 and 3 were operated such that the surface vegetation would be allowed to grow to a maximum height of about 15 cm before it would be clipped and left on the surface. In these plots, the runoff and erosion rates were less than 10% of those observed in bare plots 1 and 4.

6  THE INFLUENCE OF SLOPE AND CULTIVATION PRACTICE

The data from plots 1 and 4 can throw some light on the influence of slope on runoff and erosion. While both have bare soil surfaces, plot 4 has almost twice the slope of plot 1. It was noted that on the average, there was no significant difference between the plots in the surface runoff generated, but the erosion rates were quite different ( Fig. 3 ). The sediment yield for plot 4 was very high compared with plot 1 in the early years 1982 - 1987 (Srinivasan et al., 1988). The order of the difference reduced in subsequent years to an average of about 35% accompanied by an increase in the mean diameter of the surface soil grains, a phenomenon similar to fractional transport (Wilcock,1997).This leads to the conclusion that the silt and fine sand fractions are the most sensitive to the surface slope.

The influence of the planting method over the erosion rates can be observed from Fig. 4. Cactus was planted straight down the slope in plot 6 and along contour lines in plot 7. It was found that, on the average, both runoff and erosion were higher in the case of planting down the slope in the order of about 100 and 150% respectively. Agnihotri and Yadav(1995) found that cultivating up and down the slope, among other land uses, suppressed the infiltration rate. Thus, this practice would lead to substantial increase in runoff and erosion due to the quicker outflow as well as the suppression of infiltration rates.

7  CONCLUSION

The studies carried out in the experimental basin of Sumé, indicate that the runoff and-erosion processes even in a small experimental basin are quite complex. The varying conditions of the substrata of the soil in the region affect significantly the runoff and erosion rates from event to event. However, the large protective influence of the native vegetation against surface erosion is note worthy and any indiscriminate clearance of land for agricultural purposes may eventually lead to a total loss of surface soil and nutrients resulting in an unproductive land. The land slope affects the erosion rate much more than the runoff and the popular method of planting down the slope instead of on contoured terraces results in very high runoff and erosion. An adequate method of zoning based on the hydrologic and erosion characteristics of the basin is an essential step for a rational management of the scarce soil and water resources of the semiarid region of Brazil.

Acknowledgements

This study was initially supported jointly by ORSTOM (France) and SUDENE (Brazil) and subsequently, by the National Research council (CNPq) of Brazil through a special programme-PDCT/NE.The contributions of E. Cadier and M. Molinier in the field studies are gratefully acknowledged.

References

Agnihotri, R.C. & Yadav, R.C. Effects of different land uses on infiltration in ustifluvent soil susceptible to gully erosion. Hydrological. Sciences Journal 1995;vol.40(3): pp395-406.

Cadier, E., Freitas, B.J., & Leprun, J.C. Bacia experimental de Sumé :Instalação e primeiros resultados. (In Portuguese). Série Hidrologia no.16, SUDENE,Recife, Brazil, 1983, 89pp.

Molinier, M., Galvão, C.O. & Gomes, H.P. Determinação do escoamento de uma bacia hidrográfica através das técnicas de simulação de chuva. (In Portuguese with English summary). In: Anais, VII Simpósio Brasileiro de Hidrologia e Recursos hidricos November 17-19, 1987 Salvador, Bahia, Brazil,1987,Vol. 3:pp 530-543.

Srinivasan, V.S., Gomes, H.P., Leprun, J.C. & Silva, I.G. Erosion studies in Sumé, a semiarid region in the north east of Brazil. In: Bordas, M.P. & Walling, D, editors, IAHS Symposium, Sediment Budgets, 11-15 Dec.1988, Porto Alegre,Brazil,1988;IAHS publication No.174, pp307-314.

Srinivasan, V.S. & Galvão, C.O. Evaluation of runoff and erosion loss in micro-basins utilizing the hydrodynamic model WESP. Advances in Engineering Software 1995;Vol. 22: pp79-85. Wilcock,P.R. The components of fractional transport rate. Water Resources Research1997; vol.33(1):pp247-258.

Fig. 1  Precipitation-Runoff, Micro-Basin1.    

Fig. 2  Precipitation-Runoff, Plot No.4.

Fig. 3   Comparative Erosion in Plots 1 & 4.              Fig. 4  Comparative Erosion in Plots 6 & 7.