GrapeLook: Improving Agricultural Water Management using Satellite Earth Observation

EarthzineAgriculture, Articles, Earth Observation, Original, Sections, Themed Articles, Water Availability Theme

Image of urban flash flooding in Arlesey, U.K, 2004. Photo by Malcolm Campbell.

Annemarie Klaasse

WaterWatch BV, the Netherlands,

Caren Jarmain

University of KwaZulu-Natal, South Africa,


Limited water availability and efficient water management are major challenges facing policymakers worldwide. In South Africa, water is a critical resource for which there is strong competition between the urban, industrial and agriculture sectors. The National Water Act of 1998 states that water should be used efficiently, and has to be reserved for basic human needs and for protecting aquatic eco-systems first, with agriculture having a lesser priority. However, agriculture remains of high economic importance as it contributes to food security, export, employment and livelihood. One of the major sectors in the Western Cape Province of South Africa is the grape industry that produces wine grapes, table grapes or ‰Û÷ready to eat,’ and dried fruit or raisins.

Water authorities, together with farmers and their advisors, try to maintain agricultural production while reducing water consumption. In other words, they need to increase the water productivity. Water productivity is defined as crop production divided by amount of water consumed by the crop. It is an important indicator of agricultural performance. To improve water productivity –more crop per drop– one needs to know where and how much water is consumed: the actual evapotranspiration. It is therefore necessary to assess both biomass production and the quantity of water actually consumed by crops. The water consumption is the same as the actual evapotranspiration, which combines the vapour transpiration of vegetation and the evaporation of the soil surface. Irrigation application is not an adequate measure of water consumption as it does not include rainfall and soil moisture uptake by the plant, while losses such as runoff, seepage and percolation to the ground water are not lost from the catchment and can still be used in other areas. Actual evapotranspiration is an essential component of the water balance. In combination with the potential evapotranspiration, which is the water consumption potential for a crop without water stress, one can determine the water deficit, the amount of water the crop is missing to grow optimally.

Information on actual evapotranspiration is difficult to obtain. In situ measurements are complicated, expensive and do not show the spatial variation. The evapotranspiration is often derived from the reference evapotranspiration using universally derived crop specific crop factors. Crop factors have the drawback that they are empirically and not physically based. The single crop factor is constant over time while it can highly vary over the season. Furthermore, it does not account for water stress and provides the potential, not the actual evapotranspiration. The dual crop factor was introduced to take into account water stress and changes over time, but still does not account for differences between regions, cultivars and vineyard blocks or fields. Remote sensing based energy balance algorithms are most suited for estimating crop water use at both field and regional scales. Numerous evapotranspiration models have been developed in the last three decades using visible, near infrared and thermal infrared remote sensing data. Advanced remote sensing algorithms such as the Surface Energy Balance Algorithm for Land (Bastiaanssen et al, 1998, 2005), and the Mapping EvapoTranspiration at High Resolution and Internalized Calibration METRIC (Allen et al., 2007) have been applied to provide field level data on actual evapotranspiration and water productivity worldwide.


In 2008, WaterWatch assessed the application of remote sensing data to optimize irrigation management of vineyards in the Western Cape Province from 2004 to 2007. Farmers were very interested in the results but the retrospective nature of this study, in which the results only became available after the season ended, limited the practical application of the data on the farm level.

Image showing the study areas and provincial borders

Figure 1: Overview of the study areas (in black) with the water management areas (colored areas) and provincial borders (red boundaries).

As a result, the Western Cape Provincial Department of Agriculture, which is supported by the Department of Agriculture, Forestry and Fisheries, the Dutch Embassy and the Integrated Applications Promotion programme of the European Space Agency initiated the pre-operational demonstration service GrapeLook which was executed by WaterWatch in collaboration with the University of KwaZulu-Natal (UKZN) (Klaasse et al., 2011).

The GrapeLook project is a pre-operational service to improve water productivity and optimize fertilizer use in vineyards by providing weekly updates on crop parameters using satellite technology. The service was demonstrated for vineyards in the Western Cape Province of South Africa in the 2010-2011 season. The technology, however, is applicable at most land surfaces such as deserts, shrub lands, forests and agricultural areas.

Figure 1 shows the location of the vineyards in Western Cape Province, South Africa.

Refinement of the system took place during the demonstration phase. Information relating to crop water, growth and nitrogen status was made available freely online. The service provided weekly updates from Sept. 1, 2010, to April 30, 2011, for all major table and wine grape producing areas of the Western Cape.

Screenshot of the Grapelook website

Figure 2: The GrapeLook website.

The objectives were to:

‰Û¢ Provide weekly updated semi-real time information for individual blocks/plots and farms using satellite technology. The information included parameters such as crop growth, evapotranspiration deficits and crop nitrogen status;

‰Û¢ Forecast soil moisture change over the five days after satellite image acquisition for participating farmers only;

‰Û¢ Disseminate this information through the GrapeLook website, accessible to anyone including farmers and irrigation consultants, and;

‰Û¢ Enable farmers, water use associations, South African authorities and other users to evaluate the benefits of the pre-operational service as a tool to optimize water use and fertilizer application.


For GrapeLook, the information on the crop water and growth status is calculated by the Surface Energy Balance Algorithm for Land (SEBAL). SEBAL is an advanced-algorithm based on the energy balance. It requires inputs such as the Normalized Difference Vegetation Index (NDVI), albedo and surface temperature derived from earth observation, as well as meteorological data on the air temperature, relative humidity and wind speed. SEBAL determines actual and potential evapotranspiration on a pixel-by-pixel basis. Besides crop evapotranspiration, SEBAL estimates biomass production, evapotranspiration deficit and the biomass water use efficiency. In combination with yield data, it can be used to determine the water productivity.

Table showing Delivery parameters grouped in growth, moisture, minerals and topographyThe NDVI, albedo and surface temperature are derived from multi-spectral and thermal infrared satellite Earth observation data from the Huan Jing 1B (HJ-1B), Terra ASTER, Landsat 7ETM, Disaster Monitoring Constellation (DMC) and Fengyun sensors. The strength of the system is that it can operate by using different Earth observation satellite resources and therefore it is independent from a single source and ensures data delivery. The meteorological data in GrapeLook was derived from meteorological station measurements. The algorithm MeteoLook (Voogt, 2006) was developed at WaterWatch as a physically based regional distribution model for measured meteorological variables. It spatially interpolated weather data from meteorological stations to a raster map.

Picture of the study areas

Figure 3: Overview of study areas in the catchments, indicated with 1 (Berg River), 2 (Breede River) and 3 (Olifants River).

The map products, which include maps of biomass, water consumption and water use efficiency, were disseminated to the users through a Google Maps-based website (see Figure 2).


The dissemination website was updated weekly during the grape season of 2010-2011. Information on crop water, nitrogen and growth status at field level was made freely available online to anyone, whether working as a farmer, farmer consultant, irrigation expert or government official. The weekly parameter layers produced by SEBAL consist of 30 meter resolution raster images, which enable detailed monitoring of the temporal and spatial variations within and between blocks. Table 1 shows the complete list of parameters made available.

Satellite imagery is the major cost in the project. Although it is possible to increase the number of updates, costs would increase considerably. The idea of the project is to develop an affordable, cost-effective tool for farmers. For this reason a weekly time interval was chosen, allowing a farmer to monitor his farm and make the required changes in his management.

Picture showing actual evapotranspiration in table and wine blocks for the week of Feb 16-22.

Figure 4: Actual evapotranspiration in table and wine blocks for the week of Feb. 16-22, 2011.

The system focused on the vineyards (1) around the cities of Stellenbosch, Somerset West, Paarl, Wellington and Franschoek in the Berg River catchment; (2) around Worcester and De Doorns in the Breede River catchment; and (3) around Citrusdal, Vredendal and Klawer in the Olifants River catchment. The total Area of Interest covered by vineyards in the project extended to 1700 km2.

An example on how the data can be used, is shown in Figure 4, which presents a map of the actual evapotranspiration in an area with wine and table grape vineyards. A decision maker can make a few important observations based on this map. First of all, the actual evapotranspiration of table grapes, 50-55 millimeters per week, is larger than of wine grapes, 20-40 millimeters per week. Secondly, the water consumption in a table grape block is more uniform than in a wine grape block. Wine grape blocks can have very different water consumption patterns. A farmer can use the map of actual evapotranspiration to determine the irrigation efficiency. For example, if a farmer applies 80 millimeters of irrigation water in a week, and the actual evapotranspiration is 55 millimeters per week, it means 25 millimeters are not used by the plant during that week. Or, if two blocks receive the same amount of irrigation water but the map of actual evapotranspiration shows they consume different amounts of water, it means the farmer might reduce the irrigation application in one of the two blocks. Furthermore, the map of actual evapotranspiration helps a farmer to evaluate the effect of cultivar, soil, irrigation system and schedule, and farm management on the water consumption.

picture of  Block with wine grapes in Western Cape Province.

Figure 5: For illustration: Block with wine grapes in Western Cape Province.


The demonstration phase, which ran from September 2010 until April 2011, proved the technology could be applied on a weekly basis and that the end-users were supportive of the service. The GrapeLook solution demonstrated the potential to efficiently monitor crop water stress, crop growth, and to support better farming practices. One can expect the GrapeLook service will help reduce labor and input costs, to increase product quality and yield, and to improve water use efficiency.

In coming years, GrapeLook will continue as FruitLook. The new name reflects the inclusion of deciduous fruit trees and the extension of the target areas. Also, the website functionality will be improved to make it more user-friendly and to improve communication with farmer consultants. The existing farmer consultants and related industries play an important role as they will inform the farmers in their network about the GrapeLook service and help the farmers translate the GrapeLook data to real farm practices. For management advice, one needs to know the specific conditions of a farm and the farmer’s objective.


Allen, R.G., M. Tasumi, A. Morse, R. Trezza, J.L. Wright, W.G.M. Bastiaanssen, W. Kramber, I. Lorite and C.W. Robinson, 2007. Satellite-based energy balance for mapping evapotranspiration with internalized calibration (Metric) ‰ÛÒ applications, ASCE J. of Irrigation and Drainage Engineering 133(4): 395-406

Bastiaanssen, W.G.M., H. Pelgrum, J. Wang, Y. Ma, J. Moreno, G.J. Roerink and T. van der Wal, 1998. The Surface Energy Balance Algorithm for Land (SEBAL): Part 2 validation, J. Of Hydr. 212-213: 213-229

Bastiaanssen, W.G.M., M. Menenti, R.A. Feddes and A.A.M. Holtslag, 1998. The Surface Energy Balance Algorithm for Land (SEBAL): Part 1 formulation, J. of Hydr. 212-213: 198-212

Bastiaanssen, W.G.M, E.J.M. Noordman, H. Pelgrum, G. Davids, B.P. Thoreson, and R.G. Allen, 2005. SEBAL model with remotely sensed data to improve water-resources management under actual field conditions, J. Irrig. And Drain. Engrg. 131 (1): 85-93

Klaasse, Annemarie, Caren Jarmain, Andre Roux, Olivier Becu, and Amnon Ginati, 2011. GrapeLook: space based services to improve water use efficiency of vineyards in South Africa, 62n International Astronautical Congress, Cape Town, South Africa

Voogt, M.P. (2006) Meteolook, a physically based regional distribution model for measured meteorological variables. MSc Thesis TU Delft