Promoting Primary Production? Using Remotely Sensed Data to Investigate Human Impacts on Primary Production at a Global Scale

People are altering ecosystem form and function on a global scale through land use – changing the capacity for primary production, with consequences for the Earth System. Yet these changes are not uniform, and the interactions between population density and environmental conditions are not well established. Here we compare satellite-observed Fraction of Photosynthetically Active Radiation (FPAR) data from MODIS with a Potential Natural Vegetation (PNV) FPAR data set (Hengl, 2018), which was created using the random forest algorithm to predict vegetation properties in the absence of human alteration. Taking the average value per pixel from 2014–2017, a fixed-effects model was fitted with observed FPAR as the response variable, and predicted natural FPAR and its interactions with population density (www.worldpop.org) and biome type as the dependent variables. Population density was shown to reduce the slope of observed versus predicted FPAR, consistent with the hypothesis that the overall effect of human population density is to reduce FPAR when potential FPAR is high but to increase FPAR when potential FPAR is low. The effects differ across biomes. Maps of the difference between observed and PNV FPAR, and of the model residuals are generated to identify areas in which human activities may be promoting primary production. Whilst we limit our analysis to one of the most researched cultural variables (population density) and appreciate that our chosen data provide only a snapshot in time, with its own specific set of cultural and environmental conditions, we hope this analysis will provide a useful counterpoint to other work in unravelling human-environmental interactions at a global scale.

Hengl, T., Walsh, M. G., Sanderman, J., Wheeler, I., Harrison, S. P., & Prentice, I. C. (2018). Global mapping of potential natural vegetation: An assessment of machine learning algorithms for estimating land potential. PeerJ, 2018(8), 1–36. https://doi.org/10.7717/peerj.5457