OpenGeoHub, in partnership with the Woodwell Climate Research Center and the University of Florida, will launch a new project aimed at helping soil scientists use soil spectroscopy more efficiently and helping farmers optimize their soil management strategies.
The project, Soil Spectroscopy for the Global Good, is slated to provide a substantial increase in high quality data and analysis tools for farmers worldwide. The additional insights gained from Soil Spectroscopy for the Global Good will allow farmers to farm more efficiently and more sustainably. This could in turn enhance their food production capabilities but also benefit the environment as well as global health.
With larger quantities of food grown and harvested in more environmentally-friendly ways, a bigger proportion of the global population will have access to more and healthier food. Soil spectral analysis uses light absorbance measurements with the assistance of machine learning to examine patterns within datasets to estimate soil properties at order of magnitude lower costs than using traditional soil laboratory methods.
OpenGeoHub will contribute its own High Performance Computing infrastructure combined with functionality of Google Earth Engine and Amazon AWS for modeling. We also aim to develop methods for integrated modeling that can combine proximal and remote sensing (Sentinel-1/2 and Landsat) imagery.
As the entire system will be based on freely available open data and automated analysis, farmers, scientists, governments, and other stakeholders from all across the globe stand to benefit from the results.
To this end, OpenGeoHub is building a database with both input and output results of calibration. On top of that, we will create an open source R library with functionality for local fitting of models and a web-service for users to calibrate their soil spectroscopy readings using reliable models.
Now, low-cost soil spectroscopy equipment and access to cloud services will be all that will be necessary for participants to obtain real-time results. In the past, this process would have required a number of expensive tests lasting several days.
Soil Spectroscopy for Global Good (SoilSpec4GG):
- Jonathan Sanderman (project lead, Woodwell Climate Research)
OpenGeoHub project team involved with the SoilSpec4GG project:
- Tom Hengl (R development, spatial modeling)
- Wanderson Mendes (R development, soil spectroscopy calibration)
- Leandro Parente (HPC and web-services)
- Ichsani Wheeler (Soil data modeling and data quality control)