Analyzing the Influence of Climate Variables on Native Pasture Phenology in the Fitzroy Basin, Australia

Abstract Remote-sensing-based approaches to determining phenological parameters can help ascertain growing seasonality parameters and provide valuable insights into the growth dynamics of vegetation. Using native pastures as the vegetation type, this study aims to analyze variations and trends in the timing of the start of season (SOS) and the end of season (EOS), using Moderate Resolution Imaging Spectroradiometer data. The Fitzroy Basin in Queensland was selected as eleven-year case study, with the 2009 through 2020 period used for data acquisition and analysis. Four climatic variables—rainfall, temperature, humidity, and solar radiance data sources from climatic stations—were used to analyze the relationship between SOS and EOS. Sensitivity analysis based on multilayer perceptron neural network models was carried out to reveal the relative influence of climatic variables on SOS and EOS. Findings show that SOS and EOS trends are insignificant for most areas. Both of the phenological metrics are highly sensitive to rainfall. Fluctuation of rainfall has influence on the timing of seasonal parameters. Humidity and solar radiance, however, also appear to be a major influence for EOS. These findings can assist in providing information pertaining to pasture management strategies in anticipation of extreme climate conditions.