Evaporation models to improve water management
Local knowledge of soil evaporation for northern region soils and climates has, until now, been limited, prompting researchers to assess the way in which water loss is accounted for in prediction models. The objective is to help grain growers better manage their water use.
Globally, 20 per cent of terrestrial rain evaporates from soil, while 40 per cent is transpired through plants. Loss of water from agricultural lands through evaporation can be as high as 75 per cent of the rain that falls during a growing season.
Australian growers in particular have to deal not only with low rainfall, but also high levels of evapotranspiration – the combined evaporation and plant transpiration into the atmosphere. With drought conditions exacerbating this problem, growers are acutely aware of the need for more accurate climate and weather forecasting models.
Since 2010, researchers have recorded evaporation levels in soils in northern cropping systems using devices called lysimeters, which measure evapotranspiration. The work, led by Dr Jenny Foley at the Queensland Department of Natural Resources and Mines (DNRM), is part of ongoing GRDC-funded studies into the effects of evaporation on soil-water balance.
The data collected by Dr Foley’s team is testing current evaporation-modelling assumptions to improve the predictions of soil-water balance in models such as the Agricultural Production Systems sIMulator (APSIM), APSIM-SWIM, HowWet? and HowLeaky.
Most growers in the northern region rely on rainfall and stored soil water to grow crops, and this is factored into planting and choice of crop variety. Agronomic simulation models are often used to devise fallow management and irrigation strategies for more efficient water use.
Soil evaporation has a big influence on how much water will be stored during a fallow. It can vary considerably across different climates, soils and land management practices. Errors in modelling this evaporation can therefore have a significant impact on decision-making.
Measuring water loss
To measure soil evaporation directly, the Queensland DNRM team built a ‘lysimeter farm’ at Kingsthorpe Research Station (near Toowoomba, Queensland). The farm comprised 18 large, cylindrical, soil-weighing lysimeters. Every 15 minutes, the lysimeters measured weight gains from natural rainfall or losses from evaporation.
Researchers used four soil types cored from the northern growing region for the studies, with clay content ranging from 12 to 78 per cent. The researchers compiled a dataset for each soil after rainfall events of up to four millimetres between January 2010 and December 2013. This data covered a diversity of rainfall (light to heavy), seasonal weather variation and quantities of water available to evaporate from the soil.
Interactions between the soil water, climate and rainfall were shown to significantly influence evaporation rates.
As would be expected, all soils lose more water in the hotter months, Dr Foley says. But she says this is not only a result of warmer conditions, but also of the amount of rainfall, the amount of soil water and the way different soils lose water when evaporative demand is higher or lower.
Not all soils lose water at the same rate either. “Black vertosol [a soil from the Darling Downs] has a much higher evaporative loss throughout the year than the other soils. It loses 40 to 50mm more water per year,” Dr Foley says.
Once the research team had gathered sufficient data using the lysimeters, they were able to build this into the current prediction models to assess their performance.
Testing model prediction
Dr Foley’s team tested models including APSIM, APSIM-SoilWat, APSIM-SWIM, How Wet? and HowLeaky. Both the APSIM-SoilWat and APSIM-SWIM models were found to be accurate for modelling soil evaporation over a range of soils and seasons.
“They gave a very similar response, and were accurate in describing the measured dataset,” Dr Foley says.
Both models did show a propensity to under-predict evaporation, especially during long drying periods, but Dr Foley says this was due to inaccurate initial information rather than a deficiency in either model.
She says accurate soil physical properties and starting water states (soil water conditions) are required for the models to be as accurate as possible.
07 4529 1270
0428 309 323
All models are available via the CliMate website and app, available at: www.australianclimate.net.au
GRDC Project Code ERM00002