Our remote sensing algorithm has very high specifity and sensitivity, correctly identifying banana plantations in over 90 per cent of cases. The classifier identified 347,000 Ha of banana plantations within the geographical limits of the study. The widespread cultivation of bananas, particularly plantain species, on small-holdings or in mixed cropping systems is not captured by the mapping due to the different spectral and backscatter properties of these areas compared with the intensive areas of banana monoculture cultivation used in training the classifier. Identified banana plantations occupy a distinctive ‘climatic niche’ that differs from both the available climate space and that occupied by all croplands within the extent of the study. Plantations are located predominantly in areas of high temperature (>25C) and elevated precipitation (>2000 mm/y), with relatively low seasonality. Plantations in areas equipped for irrigation occur at higher temperatures and lower precipitation than those in areas without irrigation.
Addressing target audiences and expressing needs
- Help in technical expertise
We seek technical expertise in automating the updating of our mapping algorithm to allow real-time mapping as satellite data are updated. Our aim to to provide users with regular updates via a web-based platform.
- International Organisations (ex. OECD, FAO, UN, etc.)
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
We are preparing our manuscript for peer review publication. Following peer review and publication, we will seek partners to assist in automation of mapping.
Our algorithm can be applied to appropriate remote sensing data.
