Understanding Pastures from Space™ for South West Western Australia

Page last updated: Thursday, 19 May 2022 - 2:42pm

Please note: This content may be out of date and is currently under review.

Pastures from SpaceTM estimates green feed on offer (FOO) and the pasture growth rate (PGR). FOO is the above-ground green pasture biomass expressed as kg/ha and PGR is the current rate of pasture growth in kg/ha per day.

Pastures from SpaceTM is a joint project of DPIRD, Landgate and CSIRO and began in 2003 for the WA agricultural region.

Two services are provided on this site State level Pastures from Space map and Property Pasture Production .

Use the state level Pastures from SpaceTM map

We recommend that you read the information at Pastures from Space - FOO and PGR for agricultural properties in WA if this is your first time using FOO and PGR maps. Then open the Pastures from Space map and zoom in to your area to see current estimates of FOO and PGR. 

Use the Property Pasture Production  mapping tool

The charting function at Pastures from Space map allows you to select your property and generate property level summaries of FOO and PGR as the season progresses. It shows the median, 75 and 25% deciles calculated over all available years as well as the pasture curve for chsen years. It also allows a pixel to be selected and then displays the current years production at that level.

How does Pastures from SpaceTM estimate the FOO and PGR?

Pastures from SpaceTM estimates of FOO and PGR are based on remote sensing of 'greeness' and on-farm measurements of actual feed on offer.

The information displayed on the Pastures from SpaceTM map is supplied by MODIS (the Moderate Resolution Imaging Spectroradiometer), which is on-board both the Terra and Aqua satellites, which have been operating since 1999 and 2002 respectively (NASA MODIS). MODIS passes over Western Australia at least 14 times a week at a resolution of 250m by 250m or 6.25 ha.

Data from MODIS allows for the calculation of the normalised difference vegetation index (NDVI), which is an index between the red and near-infrared spectral bands. The NDVI is related to the photosynthetic capacity of a canopy, which is in return related to pasture biomass. The higher the NDVI, the greater the biomass.

From these 14 images, the maximum NDVI over the week is selected for each pixel to minimise cloud and other atmospheric effects. The maximum NDVI often occurs at the end of the week, but may occur earlier in the week if cloud (including shadow) makes the image unreliable. As each pixel is independent of each other and the end product is a composite NDVI image. This composite NDVI image is then used to produce the FOO image, as well as being an input into the PGR at a scale of 6.25 ha.

The Department of Primary Industries and Regional Development, CSIRO and Landgate conducted calibrations between the satellite NDVI and the FOO over a number of sites and years. From these measurements, a calibration was developed for Pastures from Space.  The PGR was compared to measurements made on farms and the model adjusted to improve the accuracy.

The PGR uses weather information (rainfall, average maximum and minimum temperature, solar radiation) over the week. This information comes from the Bureau of Meteorology weather stations. The Bureau of Meteorology interpolates the point data to develop a 5 km square grid of weather information. The PGR was compared to measurements made on farms and the model adjusted to improve the accuracy.

The interpolated weather data is then used to determine the soil moisture and temperature plant growth indexes, which both constrain growth if not in the optimum range. Land Management units also modify the PGR and they are used to take into account differences in potential growth due to soil and topographic constraints.

The last input into the PGR is the solar radiation which also comes from the interpolated 5 km square grid. The calculation uses a coefficient as an indicator of how efficient the plant intercepts light. This coefficient does vary depending on pasture composition and soil fertility. However, as MODIS cannot detect these differences a general coefficient is used which represents a clover-grass pasture.

Limitations of Pastures from SpaceTM

The relationship between NDVI and FOO is exponential therefore at the lower range of the NDVI small differences in the NDVI result in small changes in the FOO. While at higher NDVI small differences in the NDVI result in large changes in the FOO. For this reason, use of the NDVI is more accurate at FOO < 2000kg/ha. However, once FOO is more than 2000kg/ha, pasture biomass is generally not limiting livestock growth.

The predicted FOO over the growing season explains 70–75% of the variation in observed FOO. The overall mean standard error was ± 300kg/ha. There are sources of error in both the predicted and observed data so the actual error is likely to be less. The satellite NDVI can be affected by:

  • cloud and atmospheric conditions
  • differences between soil types in the NDVI signal from bare ground
  • dry pasture or stubble has a different NDVI signal compared to green biomass
  • different plants have a different NDVI value at the same biomass
  • as the season progresses there are changes in the plant structural content which adds weight but not greenness.

The relationship between predicted and observed PGR over the growing season explained 70% of the variation over three years and six farms. The overall mean standard error was ± 10kg/ha/d, which was ± 27% of the mean (38kg/ha/d). However, from the break-of-season to the end of July there was more variation with only 47% explained by the relationship. The reasons for error include:

  • when there is scattered perennial woody vegetation, low PGR values are increased and high PGR values are decreased
  • decoupling of the model occurred when low values of PGR resulted from high NDVI values. This could be due to localised weather data affecting the PGR such as frost which is not being reflected in the coarser Bureau of Meteorology weather data (5 square km grid)
  • composition differences between paddocks with different species growing at different rates.

More information

Donald GE, Gherardi SG, Edirisinghe A, Gittins SP, Henry DA and Mata G 2010, 'Using MODIS imagery, climate and soil data to estimate pasture growth rates on farms in the south-west of Western Australia', Animal Production Science, vol. 50, pp. 611–615.

Smith RCG, Adams M, Gittins S, Gherardi S, Wood D, Maier S, Stovold R, Donald G, Khohkar S & Allen A 2011, 'Near real-time Feed On Offer (FOO) from MODIS for early season grazing management of Mediterranean annual pastures', International Journal of Remote Sensing, vol. 32, no. 16, pp. 4445–4460.

Hill, MJ, Donald GE, Hyder MW and Smith RCG 2004, 'Estimation of pasture growth rate in the south west of Western Australia from AVHRR NDVI and climate data', Remote Sensing of Environment, vol. 93, pp. 528–545.

Contact information

Perry Dolling
+61 (0)8 9821 3261
Mandy Curnow
+61 (0)8 9892 8422