Ecosystem Modeling and Validation using Empirical Data from NASA CELSS and Biosphere 2

Plant productivity varies widely based on the growing conditions. Controlled Environment Agriculture (CEA) commercial production systems such as greenhouses routinely outperform open field agriculture by 10 times or more per unit area and per unit time by optimizing every resource: available light, atmospheric conditions, space, labor, nutrients and water, and by eliminating soilborne diseases, harmful pests and fungi. Dozens of mathematical and computer models of plant growth have been developed to explain and/or predict yield from growing conditions using discrete processes or functional structures, and their outputs are validated against one or more empirical studies. Validation data are typically selected based on the intended application of the model; for example, models for open-field agriculture will incorporate the range of conditions likely in the regions where a particular crop is grown and calibrated to experiments on that crop. In this study, we extend the Scalable, Interactive Model of an Off-world Community (SIMOC) with a highly generic plant growth model that incorporates 22 different plant species and validate it against two high-profile and dissimilar experiments: NASA’s Controlled Ecological Life Support System (CELSS) and the Biosphere 2 Intensive Agricultural Biome (B2-IAB). Despite a difference in yield of >10x, our model predicts the outputs of both to be within range of experimental results, and the system-level behaviors of the B2 experiment are replicated by the simulation as well. Applications of this model include holistic cost-benefit comparison of widely dissimilar agricultural practices, optimization of long-term Biological Life Support Systems (BLSS), and public education.