Profitability of climate-smart soil fertility investment varies widely across sub-Saharan Africa.

Soil fertility investments in sub-Saharan Africa, where budgetary resources are scarce, must be well targeted. Using a causal forest algorithm and an experimental maize trial dataset matched with geocoded rainfall, temperature and soils data, we modelled site-specific, ex ante distributions of yield response and economic returns to fertilizer use. Yield response to fertilizer use was found to vary with growing season temperature and precipitation and soil conditions. Fertilizer use profitability-defined as clearing a 30% internal rate of return in at least 70% of the years-was robust to growing season climate and the fertilizer-to-maize price ratio in several locations but not in roughly a quarter of the analysed area. The resulting profitability-assessment tool can support decision makers when climate conditions at planting are unknown and sheds light on the profitability determinants of different regions, which is key for effective smallholder farm productivity-enhancing strategies.

Skill (or lack thereof) of data-model fusion techniques to provide an early warning signal for an approaching tipping point.

Many coupled human-natural systems have the potential to exhibit a highly nonlinear threshold response to external forcings resulting in fast transitions to undesirable states (such as eutrophication in a lake). Often, there are considerable uncertainties that make identifying the threshold challenging. Thus, rapid learning is critical for guiding management actions to avoid abrupt transitions. Here, we adopt the shallow lake problem as a test case to compare the performance of four common data assimilation schemes to predict an approaching transition. In order to demonstrate the complex interactions between management strategies and the ability of the data assimilation schemes to predict eutrophication, we also analyze our results across two different management strategies governing phosphorus emissions into the shallow lake. The compared data assimilation schemes are: ensemble Kalman filtering (EnKF), particle filtering (PF), pre-calibration (PC), and Markov Chain Monte Carlo (MCMC) estimation. While differing in their core assumptions, each data assimilation scheme is based on Bayes' theorem and updates prior beliefs about a system based on new information. For large computational investments, EnKF, PF and MCMC show similar skill in capturing the observed phosphorus in the lake (measured as expected root mean squared prediction error). EnKF, followed by PF, displays the highest learning rates at low computational cost, thus providing a more reliable signal of an impending transition. MCMC approaches the true probability of eutrophication only after a strong signal of an impending transition emerges from the observations. Overall, we find that learning rates are greatest near regions of abrupt transitions, posing a challenge to early learning and preemptive management of systems with such abrupt transitions.