Landscape-scale predictions of future grassland conversion to cropland or development

Area of Science:

• Ecology
• Conservation Science
• Spatial Modeling

Background:

• Grassland conservation planning often overlooks landscape scale and multidimensional risk assessments.
• Existing conversion models may not align with conservation needs or ecological scales.

Purpose of the Study:

• Develop landscape-scale models to predict future grassland loss (2021-2031).
• Assess total and proportional loss of unprotected grasslands to cropland or development.
• Aid conservation planning by providing multidimensional risk assessments at relevant scales.

Main Methods:

• Developed random forest regression models for 20 US ecoregions.
• Related past grassland conversion (2011-2021) to landscape covariates.
• Predicted future loss using contemporary covariates.

Main Results:

• Grassland loss models demonstrated strong performance across ecoregions (R²: 0.83-0.96 for total loss, 0.64-0.87 for proportional loss).
• Key predictors included crop presence, distance to urban areas, ethanol plants, and animal feeding operations.
• Loss patterns varied with landscape composition; moderate (∼50%) or low (∼<30%) grassland percentages increased loss.

Conclusions:

• The developed models offer a flexible and scalable approach for creating conservation risk layers.
• These models can enhance conservation planning by enabling risk-based prioritization.
• Findings highlight the importance of landscape context and proximity to anthropogenic factors in grassland conversion.