Main content area

Are water footprints accurate enough to be useful? A case study for maize (Zea mays L.)

van der Laan, M., Jarmain, C., Bastidas-Obando, E., Annandale, J.G., Fessehazion, M., Haarhoff, D.
Agricultural water management 2019 v.213 pp. 512-520
Zea mays, case studies, corn, crop models, crop production, eddy covariance, evapotranspiration, field experimentation, irrigation, remote sensing, rivers, soil water, soil water balance, stakeholders, water footprint, South Africa
The application of water footprint accounting is mostly done at large scales, but the estimation of crop- and region-specific water footprints for up-scaling is dependent on accurate and representative in-field measurements of evapotranspiration (ET), yield and irrigation. In a field trial we assessed the influence of maize (Zea mays L.) ET estimates using soil water balance accounting, remote sensing with satellite imagery (SEBAL model), eddy covariance measurements and three crop models (SWB, CROPWAT, SAPWAT) on water footprint estimates. We simultaneously assessed the influence of yield spatial variability as measured by a precision harvester. Seasonal ET estimations differed by as much as 15% for the different methods, and yield differed by as much as 42%, representing the error which can be introduced as a result of point measurements. Using a combination of the highest/lowest ET estimates and the 5th/95th percentile yield, water footprint values differed by as much as 100%, ranging from 338-680 m3 t−1. Applying spatially-linked SEBAL ET estimates and precision harvester yield at the 30 × 30 m scale reduced the range of estimated water footprints to 493-663 m3 t -1, with an average of 547 m3 t-1. This was 15% higher than the water footprint estimated using average SEBAL ET and average yield for the whole pivot (467 m3 t -1). Any error introduced at this stage of water footprint accounting can be transferred during up-scaling of the results. For example, based on the minimum and maximum estimated water footprints, maize production was expected to consume between 4.4 and 8.3%, respectively, of the Orange River (South Africa’s largest river) flow during the season in that region. Biophysical scientists have the role of providing high quality data for accurate water consumption estimates. Thereafter, their application by various stakeholders should be done with caution.