Remote sensing evapotranspiration in plantation forests
The development of the remote sensing model for estimating evapotranspiration in the Lower Limestone Coast’s plantation forests. Using advanced machine learning techniques, the model provides accurate, high-resolution evapotranspiration (ET) data critical for sustainable water management in the region.
The challenge
The Lower Limestone Coast faces significant water management challenges, particularly in balancing the needs of large-scale plantation forests and groundwater resources. Accurate estimation of evapotranspiration (ET) is crucial for understanding the water consumption of tree plantations like Pinus radiata and Eucalyptus globulous. Traditional methods have proven insufficient, often leading to inaccurate water usage assessments. By developing the remote sensing ET model, this project provides precise, 30m-resolution ET data that enables better decision-making for water sustainability. This work is essential for the region’s long-term environmental health and economic productivity, ensuring responsible forest management and groundwater preservation.
Field measurement of ET in the Lower Limestone Coast region. The red circles show the sites that were used. The base map is the average annual Landsat 8 NDVI from 2015 to 2020 where lighter colours show increasing landscape greenness.
Our response
To address the challenge of accurate water usage assessment in the Lower Limestone Coast, we developed the AMLETT (Australia-wide Machine Learning Evapotranspiration for Trees) model. This innovative approach combines field data with remote sensing technologies, including Landsat 8 and vegetation indices like NDVI and EVI, to estimate evapotranspiration (ET) at a 30m resolution. By employing Random Forest machine learning techniques, the model accounts for seasonal variations and specific characteristics of plantation forests such as Pinus radiata and Eucalyptus globulous. The result is a highly accurate ET dataset that improves on previous models like CMRSET, especially in areas where traditional methods have over- or underestimated water consumption. This project is a collaborative effort involving key partners. The Commonwealth Scientific and Industrial Research Organisation (CSIRO) leads the research, with support from the South Australian Department for Environment and Water.
Predicted annual ET from 2000 to 2020 for the Limestone Coast plantation region.
The results
The expected outcome of this project is the creation of a highly accurate evapotranspiration (ET) dataset specific to plantation forests in the Lower Limestone Coast. By applying the AMLETT model, water managers will gain precise insights into plantation water use, allowing for better management of groundwater resources. This will lead to improved decision-making, helping balance the region’s environmental and economic needs. The model also provides a scalable framework for future ET estimations in other areas, contributing to broader efforts in sustainable water resource management across Australia.
Find out more
Click here to see the full project report: http://hdl.handle.net/102.100.100/602968?index=1