|Management of human-induced salinisation in the Berg River catchment and development of criteria for regulating agricultural land use in terms of salt generating capacity
|Expanded Title:||EXECUTIVE SUMMARY
From previous WRC work by the same group, various needs were identified. One was to examine the effects a range of land uses may have on the production of salinity from the Sandspruit catchment. Another was to develop criteria to manage the salt production from this area. In the days of Dr Martin Fourie (1976), it was apparent that the load of salinity reaching the Berg River differed from sub-catchment to sub-catchment along the river. This project set out to establish if the range of agricultural practices in the region could be related to the differences in salt production in any way. In so doing, the team had to analyse the whole drainage path towards the river, make sure that this was duly represented in the models that were used and then test the effect that land use change had on the salinity and water levels that reached the Berg River. By implication the study did not focus on hydrological modelling only, but also on the understanding of the system through research aimed at determining the origin of salt, the behaviour of soils in this landscape and the effect different cultivation practices had on the water flow path. Subsequently, it had to be ascertained whether the hydrological models assessed these flows in an appropriate manner.
The main aims and the specific objectives of the proposed project were:
• Quantification of the water and salt balances for a variety of land uses and on-farm management practices.
• Setting up and development of a hydrological model of the Sandspruit catchment for predictions of salt load contributions to the Berg River from different land uses and management practices, and finding the best practice to accommodate land use change in hydrological modelling.
• Development of guidelines for regulating land use in terms of salt generation capacity, based on the knowledge gained from on-farm experiments and hydrological modelling.
Continuation of monitoring since the project by De Clercq et al. (2010) helped to establish fairly long-term trends in salinity, soil water and ground water trends for this region. These were necessary to establish a pattern of behaviour related to specific land use practices. Defining the land use differences and land use change were therefore done by describing the differences in water movement through the different land use sites indicated for specific regions in this study. The sites were as follows:
1. The Voëlvlei area, where a Renosterveld stand was compared to an adjacent wheat land in a two year crop rotation cycle under conventional till.
2. The Goedertrou farm, which is a long-term monitoring site, with a crop rotation cycle of 3 years, and a young site in terms of minimum tillage.
3. The Langgewens experimental farm, where there is a range of cultivation practices, some as old as 30 years. Here the cultivation and crop rotation practices were also monitored in terms of soil-water content, overland flow, infiltration rates and changes in soil density.
4. The Sandspruit catchment, which was used to monitor borehole information related to water table depth responses, salinity and characterisation of concentration levels of selected isotopes.
A direct comparison in terms of water relations was done on the Voëlvlei/Schoongezicht site between the Renosterveld and a wheat production system. They were compared in respect of their seasonal soil water status, the perched water table levels and the deep water tables. The data were modelled using Hydrus-1D.
Specific land use scenarios were also studied on the Langgewens experimental farm where the impacts of cultivation practices on the water flowpaths were investigated.
RESULTS AND DISCUSSION
The results of the investigation are shown in the following figures. From these figures, it can be seen that the water usage (evapotranspiration) in the Renosterveld system by far exceeded that of the wheat system. This explains the lower water table experienced in the Renosterveld when compared to that in the Wheatfield regions.
Tracer studies were done on the Goedertrou farm and in the Sandspruit catchment. On Goedertrou, the idea was to test the role which salts in the soil played in the salt production from an area. During 2005, potassium iodide (KI) was introduced on the farm in an infiltration study. During 2012, the locations where KI was introduced were re-visited and the reductions in iodine at these locations were measured. It was found that the iodine, which resembles chlorine to a large extent in terms of behaviour and reactions, was reduced significantly and the manner in which the reduction took place is highly significant to this project. We found as follows:
1. The iodine in the normal minimum tillage environment decreased the least in the upper 5 cm of the B-horizon. Below the B-horizon very little was left – and likewise in the A-horizon.
2. Where we introduced KI behind the contour bank, no iodine could be found.
3. We tested for iodine in the nearby farm dam and traces of iodine could be found in the evaporite formed in the dam sediments.
4. The tracer study conducted in the Sandspruit catchment confirmed aspects of water movement through the system.
In the Sandspruit all results indicated that groundwater is possibly affected to a larger extent by in situ deep drainage than by lateral flow in the system.
Langgewens cultivation practice
In the Langgewens cultivation practice the study was performed on an old but still continuing experiment where four different tillage practices were applied since 1975. Of these, conventional till and no-till were the most extreme in terms of rain water uptake. The study therefore focused mainly on these two practices. It was indicated that the rooting depth, soil salinity, soil structure and soil density were all affected negatively by conventional tillage practices, while with no- and minimum till all these factors tended to return to their natural state. The no- and minimum till practices therefore presented soils with lower salinity, deeper rooting volumes and less compaction, generating conditions that allowed faster water uptake and deeper infiltration of rainwater. These sites also indicated a lower risk for the farmer during dry spells in winter. The major concern is that the better practice therefore seems to be that where less tillage is practiced. Less disturbance of the soil implies more winter water that enters the soils and feeds into the water table. Larger disturbance of the soil implies more overland flow. Our concern therefore is that the agricultural practices of no- or minimum till have been shown to be the preferred practice for farmers and conservation farming. In terms of water and salinity in the catchment, it is apparent that more infiltration into the soils also implies more winter water that enters the system. This presented a larger saline seep problem until late in summer. The extra winter water in these systems is therefore not utilised by the crop production system and the removal of unwanted (alien or indigenous) perennials is therefore not advisable if saline seep needs to be limited.
Langgewens overland flow
The Langgewens overland flow and hill slope study was done measuring overland flow in the crop production landscapes on the farm looking at a range of terrain types (Figure 4). For all the measuring points, the ratios between surface runoff, sub-surface flow and deep drainage was monitored. It was indicated that a considerable component of the water reached the water table as deep drainage. These responses were modelled using Hydrus-1D.
Remote sensing of salinity
Remote sensing of salinity was done to take stock of the salinity in the catchment. It was also done to develop methods through which better input could be generated for hydrological and salinity modelling. These results further had to contribute to our understanding of the surface processes and to the defining of the Hydrological Response Units (HRUs) in this area. The occurrence of salinity in the catchment was therefore mapped based on measurements and the team’s experience in terms of where soil salinity occurs. This can be seen in Figure 5. Red indicates areas likely to be affected by salinity and green shows areas which are less likely to be affected by salinity based on Digital Elevation Model (DEM) derived curvature.
Saline prone areas were mapped based on a 20 m DEM and borehole data (Figure 6). The salinity map was derived from an electrical conductivity range predictive model based on a DEM. Blue areas are less likely to be saline, yellowish areas are likely low to moderately saline and red areas are strong to extremely saline.
The hydrology of the Sandspruit was tested in a number of model set-ups.
Quantification of the water balances
The hydrology of the Sandspruit was tested with the J2000/JAMS model using a number of model set-up scenarios. Land use and land use change for the Sandspruit catchment are extensively linked to winter crop production systems. The measurements and models were therefore prepared to be sensitive to all possible situations in a complex matrix of crop rotation in the catchment. Testing of the models was therefore done in a real life situation and extremes where all land use in the catchment was modelled for grazing and a total wheat cover.
The land use was mapped using satellite data. The reason for using satellite data were that one could utilise the Landsat database which spans the last 30 years. However, the development of historic database of land use from Landsat information fell outside the scope of the research. An analysis of the effect scale had on the parameterisation of a model for the Sandspruit catchment, brought to light that running a model for the Sandspruit scale, implied that specific variation in soil physical parameters and small changes in land use related to the scale of the elevation model, did not affect the model significantly enough.
At first it was found that the model could not simulate the expected salinity response at the Sandspruit weir accurately enough. The weir response was slower than the model predicted response and the salinity prediction at the weir was equally distorted. The salinity routine was therefore adapted to reflect our understanding of the salt transfer and the reason for the earlier response was found in the way the contour banks influenced the overland flow. A new routine was developed and implemented in the model to make sure the contour banks affect the runoff in the correct manner.
Setting up of hydrological models
Setting up of the J2000/JAMS model was done implementing a new salinity routine and implementing the first contour bank routine used in any hydrological model of the world. For the benefit of this research, the model was calibrated using a time span for which good data existed. This calibration was used to simulate the balance of the available data to calibrate the model. After calibration, the model was used to predict the normal mix of land uses, a change in land use towards a single land use of grazing only, and a single land use of wheat only. The effect which these three scenarios have on runoff of water and NaCl are indicated below in Figures 7 to 9.
Guidelines for regulating land use were therefore defined by five major criteria, namely:
1. The crop area, and most importantly the total surface area, covered by perennials and annuals and the ratio of these areas to the total landscape area of a region.
2. The cultivation practice – affecting the soil and water flow paths and the contribution to stored water in the system.
3. The presence of contour banks and their role in water flow paths, as well as their role in water retention (slowing of runoff), thereby increasing deep drainage and recharge.
4. The occurrence of salinity in the landscape.
5. Accumulation of water in the landscape and the routes water use to reach the river.
The project achieved more than the aims set for this research programme. The models chosen for the purpose of predicting water and salinity during land use change were even more applicable after the models were adapted. The J2000 model was adapted to be a useful tool in the evaluation of land use practices for this region. The lessons learnt will also help us to improve the ACRU model. The Hydrus-1D model was used to help us understand the accumulation of water in the landscape as well as the role played by cultivation practices.
Regarding the hydrological model input it was necessary to make absolutely certain that the models we used would be sensitive enough to be able to reflect the effects of land use change in the differences between measured and predicted runoff from the Sandspruit. This research consequently focused largely on understanding of the systems and to make certain that the modelling output could be presented with confidence.
A broad understanding of how these agricultural systems operate and the influence on the environment was developed. This understanding, which includes the soils, the crops, the weathering zone, the water table, amongst others, guided the modelling done for the research.
The following key findings were indicated by this research:
1. That land use, with specific emphasis on cultivation practices, determines the ratio between water infiltration and overland flow, also the subsurface lateral flow response and consequently the amount of water that will reach the water table.
2. That land use is in fact a key factor to salt and water responses in this environment.
3. That minimum tillage and no-till increased the total water and salt response measured in this region.
4. That it is possible for farmers to contribute to the reduction of summer saline seep that reaches the Berg River.
5. That the use of contour banks needs to be considered in modelling.
6. That the crop production systems, compared to perennial plants, indicated that low utilisation of water in the crop production systems, is the major cause of the saline seep problem.
7. That through hydrological modelling we can design agricultural land use practices that could assist salinity management in this region so that the negative impact of dry-land salinity on our fresh water resources is lowered.
8. That deep-rooted natural vegetation in the major drainage pathways could stop summer seep sufficiently to reduce or stop saline seep from the low lands to reach the river.
This research will not stop the region from being a crop production area. The knowledge gained from this research could however change the modus operandi of the Working for Water programme for this region. We cannot condone the removal of trees and shrubs to allow more water to reach the streams. This results in elevated saline water tables that also have a much worse environmental impact.
Since the removal of trees and shrubs allow more water, and accompanying salt load, to reach the streams, this practice needs to be suspended in this region.
We recommend that future research focus on the alternative farming practices that could lower the summer seep towards a river system. Research should therefore also be focussed on the re-vegetating riparian zones with adequate indigenous deep-rooted plants.
We also need to focus on the implementation of management tools for farms and catchment managers.
The long-term monitoring of the current research infrastructure should be maintained. The total set of boreholes were sunk with a total cost of close to R4 m and could over time become increasingly important in our quest to adapt to the effects of climate change. There are significant advantages in having sites like these for use in global change research.
Recommendations for salinity management
The lessons learnt during this study are quite simple. We do have a typical dryland salinity situation in the lower Berg River catchment. The management of salinity in the landscapes is necessary and can be achieved by utilising the excess winter water in the crop production systems. Too much water is stored resulting in the elevation of water table and salinity in the regolith, causing saline seep until late in summer. The management practice should be to utilise the excess winter water with more perennial deep-rooted halophytic vegetation.
Monitoring of the perched winter water tables should be done and the utilisation of deep-rooted vegetation in riparian zones should be the norm. Furthermore, buffer zones should be used along all waterways, even sporadic streams, to enhance the impact riparian zones could have on the management of saline seep.
It is strongly recommended that the findings of this research be developed into workable management practices. It is further recommended to adapt the Working for Water programme for regions where dryland salinity occurs.
|Document Type:||Research Report
|Document Subjects:||Agricultural Water - Small holder irrigation
|Document Keywords:||Water Quality
|Document File Type:||pdf
|Research Report Type:||Standard
|WRC Report No:||1849/1/13
|Authors:||de Clercq W; Jovanovic N; Bugan R; Mashimbye D; du Toit T; van Niekerk A; Ellis F; Wasserfall N; Botha P; Steudels T; Helmschrot J; Flugel W
|Organizations:||Stellenbosch University; CSIR; Friedrich Schiller University, Jena, Germany
|Document Size:||15 255 KB