The COALA project has been operational since January 2020. Our project uses information provided by satellites (the Sentinels) to help support Australian farmers in the Murray-Darling Basin. One of the ways that we support farmers is through precision irrigation. Precision irrigation is a farm management approach that involves using technology to make informed decisions.

The Challenge

Significant progress has been made to utilise precision irrigation worldwide. Precision irrigation increases water use efficiency or decreases the water footprint in irrigated agriculture. The progress is mainly restricted to advances at small scale. The information that new satellite platforms provide are rarely fully exploited and made available to farmers.

COALA aims to provide farmers and water managers with products based on Earth Observation. These products are state of art spatially distributed physically-based which estimate water fluxes (crop water requirements, evapotranspiration, soil water content, irrigated areas) from district scale to farm scale. In other words, COALA is developing innovative multi-scale information products shared by the different levels of users.

How to exploit earth observation images for water management beyond the state of art?

The data products and services offered by COALA are based on a combination of Earth Observation data provided by COPERNICUS Sentinel satellites and complementary data derived from different data sources.

Monitoring the crop development in effective way, beyond the classical Vegetation Index approach.

The monitoring of crop vigour is based on the widely used Normalized Difference Vegetation Index (NDVI). Indeed, COALA provides time series of parameter easily recognizable in the fields. For example, the Leaf Area Index represents the square meter of leaf surface per square meter of the soil surface. Leaf Area Index is retrieved using Neural Networks Algorithms by exploiting all the information content of the new Sentinel satellites.

Improvement in the identification of irrigated paddocks.

COALA develops multitemporal approaches and machine learning algorithms to detect irrigated areas with high accuracy levels. We identify the irrigated and not irrigated paddocks using several different Machine Learning Algorithms. These Algorithims are based on dense time series of vegetation indices. The indices exploit all the bands of the multispectral satellite acquisition (visible, near and shortwave infrared).

More accurate estimation of actual crop evapotranspiration

Evapotranspiration is the key variable for determining crop water requirements. Existing services estimate the evapotranspiration based on the crop model and well-known Kc and NDVI relationship. IrriSAT Australia currently uses this method. This method has several critical points, mainly because the relationship is based on subjective evaluations of crop development in the field. There is no effective way to estimate its accuracy.

COALA introduces an innovative approach for estimating crop evapotranspiration. We exploit the full capabilities of the Sentinel-2 platforms in terms of geometrical, temporal and radiometric resolutions. It is based on a modified combination equation FAO-56 for computing evapotranspiration by incorporating shortwave infrared data from Sentinel-2 to assess the surface’s water status and modulate the resistance terms in the model. Technically, shortwave infrared data permits a more accurate evaluation of crop resistance in soil water limiting conditions.

Estimated turnaround time

Sentinel-2 satellites acquire a new scene every 5 days. After the satellite passage, the related products (Leaf area index, NDVI, soil water status, fractional cover, etc.) are available usually the day after the passage. Then, evapotranspiration (ET), crop and irrigation water requirement (CWR, IWR) are calculated based on the last satellite acquisition and meteorological data forecast. For this reason, they are available every day up to date with a forecast of up to 5-10 days.