Skip to content
Start of page content below the header

Reporting from the Agri4D Conference: Water Resources and Sustainable Intensification

Global forecast for 2050 is a continued growing, increasingly urbanized and more affluent population, which, most likely, will demand both more food and more water intensive food. As a result, the pressure on already stressed water resources for food production will increase further.

The major part of future production and consumption of food will take place in developing countries where water availability often is a limiting factor. Consequently, a large part of the future food demand must be met through “sustainable intensification” of agriculture. Well-chosen management of precious land and water resources will be crucial.

Through four presentations this session highlighted linkages, modeling approaches, new perspectives, options and trade-offs of how to best orchestrate agricultural water management across scales to achieve sustainable intensification of food production.

The first presentation Water, forests and footprints – finding the right scale for sustainability” was given by Kevin Bishop, Professor at both the Department of Earth Sciences at Uppsala University, and at the Department of Aquatic Sciences and Assessment at the Swedish University of Agricultural Sciences (SLU).

In his presentation Kevin Bishop tried to unfold how forests impact water partitioning at different scales. At a local catchment scale, a removal of forests usually increases the total flow, always increases the peak flows, and can increase as well as decrease the base flow. Although there exists some ambiguity regarding the role of forests, all scientific studies confirm that forests have larger evapotranspiration (ET) than most other land uses.  (Evapotranspiration is the sum of evaporation and plant transpiration from the Earth’s land surface to atmosphere) However, there are studies that consider impacts of forests on water availability for annual mass balances at the local watershed level misleading.

When forest ET is viewed at larger spatial scales and at narrower seasonal temporal scales, it is possible to shift the perspective from a local “demands” towards a regional “supply” viewpoint. Thanks to the high ET from forests more precipitation is generated over the continents within the same year, increasing the seasonal rainfall. Thus, while afforestation, or reforestation, might lead to reduced runoff at the local catchment level, i.e. the stream flow no longer meets downstream demands. The increased number of trees might at the same time increase the seasonal precipitation at the larger regional scale, i.e. increasing the supply.

In conclusion, forest cover plays an important role for the hydrologic cycle. On a regional scale an increased forest cover can increase the overall precipitation and runoff. Consequently, local decision-making about forests will have trans-boundary impacts.

After that Dr. Jafet Andersson at the Meteorological and Hydrological Institute (SMHI) presented his study on “Computational approaches to address water resource challenges and agricultural development – examples from Africa, India and Europe”. Through a number of regional examples Jafet Anderson showed how computer models can assist in providing quantitative information about how different interventions to increase crop yields play out at different scales. By using spatiotemporally dynamic integration of multiple processes, feedbacks and scales it is possible to make systematic assessments. Estimates can also be given in unmonitored areas.

Jafet Anderson used SWAT modeling, Soil and Water Assesment Tool, to show how water and nutrient interventions might impact crop yields in South Africa and the Tukela River basin. The primary technologies were, a) in situ water harvesting, b) external water harvesting, and c) human urine fertilization (Ecosan), including scenarios with d) full irrigation and e) full fertilization. The studies show that the most important limiting factor for crop yields is the lack of nutrients, and limited water availability.

Current work at SMHI focuses on analyses of how climate change will impact local water resource availability. Impacts are estimated by combining global climate modeling (GCM), dynamic regional downscaling (RCM), bias correction and scaling (DBS) and the hydrological model HYPE. One example of ongoing work is the impacts on irrigation water demands in large-scale agricultural systems in Europe and India.

Although modeling can give many answers, it is often necessary to include participation by local stakeholders. SMHI has ongoing computation and capacity development projects in Niger, Botswana and the MENA region. The studies focus on how climate change impacts frequency and magnitude of floods and droughts, the agricultural performance and energy generation.

Through open models and data access the model development and analysis can become a shared process. The aim is to generate local ownership of the process in order to eventually facilitate sustainable management and adaptations. In addition, following strong initial involvement by local stakeholders future refinements and application will be possible without external actors

In conclusion, computational approaches can be useful to assess challenges and opportunities for water resources and agricultural development (South Africa, India, Europe, and West Africa). However, modeling is not enough. Success depends  on people, data, policy, institutions etc.

Dr. Louise Karlberg and PhD student Yihun Dile, Stockholm Environment Institute (SEI), Stockholm, and the Stockholm Resilience Centre (SRC), Stockholm University presented their work on “Agricultural water interventions for sustainable intensification – upstream downstream trade-offs and opportunities” This talk presented two sister projects in Ethiopia and India. In both case studies the SWAT model was used to analyze how scenarios of upstream water harvesting and nutrient application interventions impact downstream water availability.

The case study in Ethiopia shows that crop yields significantly increase with water harvesting and nutrient applications. By only implementing water harvesting yield scenarios show an increase by 65 % and by adding nutrient applications yields improved by up to 200 %. Water productivity also increases with water harvesting and application of nutrients.  However, there is  upstream-downstream water availability trade-offs that need to be take into account. The results show that total stream flow decreases after implementation of water harvesting. The frequency and magnitude of stream low-flows has increased, which implies improved water availability in the dry season.

The sister project in India also incorporates differences in farm income before and after implementation of water harvesting. As a result of the project, most farmers have shifted from cropping traditional staple crops, like sorghum, to cotton, which is a cash crop. With improved water availability during the dry season many farmers have also started to grow vegetables for sale on the market. Thus, farmers have moved from subsistence farming to a more commercial agriculture. In dry years the difference in farm income is very small. However, during “normal” agricultural years, which most of the years are, farm income nearly doubled with the “new” agricultural system. Looking at water balance changes, the ground water recharge increased for all years, while the outflow from the catchment in dry years was reduced by up to 60 %. This reduction in runoff impacts the drinking water availability for downstream Hyderabad, the Karnataka state capital. In addition, results show a significant reduction in sediment loss.

Both studies indicate water harvesting has significantly decreased runoff. Further, peak flows are reduced and low flows increased, thus reducing the risk of flooding and erosion. Sediment losses were also reduced after water harvesting implementation. In upstream areas crop yields and biomass production have increased. In downstream areas water availability for drinking water supply may be reduced, as was shown in the study in India. The final presentation “Precipitation-sheds and the resilience of green water systems – defining a new set of challenges for landscape management” was given by Ph.D. student Patrick Keys at the Stockholm Resilience Centre (SRC), Stockholm University and the Department of Atmospheric Sciences, Colorado State University.

This presentation introduced the audience to precipitation sheds, defined as places where precipitation originates, and the importance, and difficulty of setting the boundaries of such regions. Precipitation-sheds are very dynamic in their nature with large variability in both time and space.

Mr. Keys presented literature review of the effects of anthropogenic land use changes on downwind areas. Results show several examples where strong signals are found for patterns of upwind – downwind connections. For example irrigation upwind causes increased precipitation downwind in many regions in India, China, and the United Stated. He further presented possible management strategies for precipitation sheds, and limitations of such methods. For  instance, Mr. Keys points out that the area for a precipitation shed is huge, thus very difficult, but not futile, to manage,

In summary, it may be important to monitor land use changes in certain upwind areas since changes in evapotranspiration may have large impacts on precipitation, and thus on moisture availability for crop production in downwind areas. Hence, research on precipitation-sheds may identify places and patterns of particular importance and possible threats to such patterns, e.g. global warming.

The closing session was a panel discussion between all presenters, chaired by Dr. Mats Lannerstad at Stockholm Environment Institute (SEI) and International Livestock Research Institute (ILRI), Nairobi. Questions from the audience mainly were focused on how modeling approaches can be improved and how the results can be used locally.

In particular, the importance of well defined boundaries and concepts were discussed. The panelists highlighted the importance, but also the complexity of  integrating social aspects in their models. Further the complicated ownership of water resources, especially green water resources, was brought to the surface, with the conclusion that access to information and knowledge sharing will assist sustainable usage of water resources.

Last words regarding future challenges from the panelists:

Jafet Anderson: We have been compartmentalized? What about policies that influence agriculture? We are in an integrated society. This is a fundamental challenge and a key focus is to integrate many more processes in our models. What we can do depends on what scale we are looking at.

Patrick Keys: The key challenge is to bring my work, which is at a very large scale, to a applicable scale, to develop something that can actually be used by people and and integrated in their life.

Yihun Dile: My instant reflection is on resilience. With water harvesting we can have sustainable agriculture, which is resilient. But, how can people pick up the responses?

Louise Karlberg: Previously, We had undesirable resilience. We don’t want to have subsistence agriculture. We are happy that the resilience was low and the shift to commercial agriculture could be adopted. What can we say about economic tipping points?

Rapporteur Ylva Ran, Research Associate, SEI