Artificial intelligence and big data analytics have been applied to produce the first global map of the world’s regions where vegetation can and cannot be grown.
The Valencia University study assesses the global abundance of the phosphorus and nitrogen content in vegetation. Also assessed is the efficiency in water use. The scientists’ aim is to show where the best places are for agriculture and where environmental conditions are changing in response to climate change. The application of artificial intelligence and big data methodologies also enables an assessment to be made of our planet’s biodiversity.
Together with carbon, hydrogen, oxygen and sulfur, nitrogen and phosphorus are the principal chemical elements incorporated into living systems. They are strong signals of the suitability of different parts of the Earth for agriculture. Both nitrogen and phosphorus are needed by plants in large amounts (although excessive quantities can also cause environmental damage). In soil, nitrogen and phosphorus are typically found in the form of nitrates and phosphates.
The new global maps produced by the researchers gathered information from Google mass satellite observation data and then used a specially developed artificial intelligence program to assess the data and produce the color-coded maps. The satellites gathered temporal and spatial observations, and this produced a series of maps characterizing different biophysical parameters. To develop the maps required numerous observation-measurement pairings to be number crunched.
Speaking with Phys.org, lead researcher Álvaro Moreno explained why the maps were significant: “Until now, it was impossible to produce these maps because the required conditions weren’t available. We didn’t have powerful and accurate machine learning statistical tools, nor did we have access to great bodies of data or cloud computing.”
The new maps and the process behind them are published in the journal Remote Sensing, in a paper titled “Regional Crop Gross Primary Productivity and Yield Estimation Using Fused Landsat-MODIS Data” and an companion article in Remote Sensing of Environment titled “A methodology to derive global maps of leaf traits using remote sensing and climate data.”
The next steps are to use the technology to further assess the impact of climate change and to assess other important societal and ecological questions like the pressure on food production to meet population growth and the development of new technologies, like biofuel production.
