Conference Publication Details
Mandatory Fields
Marwaha R.;Cawkwell F.;Hennessy D.;Green S.
Precision Livestock Farming 2019 - Papers Presented at the 9th European Conference on Precision Livestock Farming, ECPLF 2019
Improved estimation of grassland biomass using machine learning and satellite data
2019
January
Validated
1
()
Optional Fields
ANFIS Grassland Landsat-8 Machine-learning biomass
174
179
Precision Livestock Farming 2019 - Papers Presented at the 9th European Conference on Precision Livestock Farming, ECPLF 2019. All rights reserved. Monitoring management practices in grasslands such as grazing and silage cutting at field scale helps to understand the yield and its carrying capacity. Traditional methods for grassland monitoring can be time consuming and labour-intensive. In this study we estimate grassland biomass at a farm-level using remote-sensing. For a full understanding it is important to monitor the grass growth and utilisation on the farm with a high temporal and spatial resolution satellite images which can be used for monitoring biomass, phenology and growth rate of a pasture. The main goal is to determine grass growth rate using satellite and weather data (Moorpark, Co. Cork in Ireland is used as an example). Normalised difference vegetation index (NDVI) from Landsat-8 satellite, along with weather data such as temperature, rainfall and potential evapotranspiration data were used to model grass yield and compared with ground measurements and traditional biophysical grass growth model outputs. Adaptive neuro-fuzzy inference system (ANFIS) was used to generate models to predict and map grass biomass. The models were evaluated using Mean Squared Error, Root Mean Squared Error (RMSE) and Symmetric Mean Absolute Percent Error (SMAPE). The final model was used to predict the biomass for the subsequent year for validation purpose. The resulting grass growth rate will help a farmer in understanding the maximum potential from their farm as well as the stocking rate on the farm.
Grant Details