OOS 9-3
Global crowdsourcing of soil-specific woody plant encroachment patterns and management strategies
The rate, extent and persistence of woody plant encroachment vary at hectare to global scales. Our ability to manage encroachment is constrained by limited knowledge of spatially- and temporally-specific factors. Experimental approaches are insufficient because they capture neither the range of variability in site conditions, nor the diversity of management responses that have been and are being tested by land managers. Even when documentation of encroachment and successful responses are recorded, documentation rarely includes sufficient information on site conditions to determine relevance to other locations in the region or globally. A Land-Potential Knowledge System (LandPKS) is under development that will allow global crowdsourcing of site-specific information and management history. The information will be provided as user inputs by individuals interested in knowing potential land productivity and resilience. Core (required) inputs will be entered using a user-friendly app and will include basic soil profile, topographic and land cover information. The system is being designed modularly so that additional information can be added, including more detailed information on land cover, management and other disturbances. This information will be integrated with cloud-based geospatial climate and other data to provide an assessment, while contributing to a soil-specific global database.
Results/Conclusions
Testing of an early beta version of the app in Kenya and Namibia indicates that the input system works relatively well for basic site information. An early version of an integrated soil color tool that uses the phone’s camera and a color reference card generates relatively consistent predictions for sieved dry soil, but is less accurate and precise under field conditions. Accuracy of the soil texture key, which uses a combination of images and short video clips, varies with user knowledge and experience. Enhancement of both tools continues. The app is being designed using icons to minimize or, with training, potentially eliminate the need for the use of written language for input. While we have had some success with this approach for basic site information documentation of management history is more challenging. We are currently exploring additional options for addressing the diversity of global languages and, within languages, local terminology for describing specific management practices in cooperation with a large number of partners.