Identifying links between monsoon variability and rice production in India through machine learning.

Climate change poses a major threat to global food security. Agricultural systems that rely on monsoon rainfall are especially vulnerable to changes in climate variability. This paper uses machine learning to deepen understanding of how monsoon variability impacts agricultural productivity. We demonstrate that random forest modelling is effective in representing rice production variability in response to monsoon weather variability. Our random forest modelling found monsoon weather predictors explain similar levels of detrended anomaly variation in both rice yield (33%) and area harvested (35%). The role of weather in explaining harvested rice area highlights that production area changes are an important pathway through which weather extremes impact agricultural productivity, which may exacerbate losses that occur through changes in per-area yields. We find that downwelling shortwave radiation flux is the most important weather variable in explaining variation in yield anomalies, with proportion of area under irrigation being the most important predictor overall. Machine learning modelling is capable of representing crop-climate variability in monsoonal agriculture and reveals additional information compared to traditional parametric models. For example, non-linear yield and area responses of irrigation, monsoon onset and season length all match biophysical expectations. Overall, we find that random forest modelling can reveal complex non-linearities and interactions between climate and rice production variability.

Crop yield response to climate change varies with cropping intensity.

Projections of the response of crop yield to climate change at different spatial scales are known to vary. However, understanding of the causes of systematic differences across scale is limited. Here, we hypothesize that heterogeneous cropping intensity is one source of scale dependency. Analysis of observed global data and regional crop modelling demonstrate that areas of high vs. low cropping intensity can have systematically different yields, in both observations and simulations. Analysis of global crop data suggests that heterogeneity in cropping intensity is a likely source of scale dependency for a number of crops across the globe. Further crop modelling and a meta-analysis of projected tropical maize yields are used to assess the implications for climate change assessments. The results show that scale dependency is a potential source of systematic bias. We conclude that spatially comprehensive assessments of climate impacts based on yield alone, without accounting for cropping intensity, are prone to systematic overestimation of climate impacts. The findings therefore suggest a need for greater attention to crop suitability and land use change when assessing the impacts of climate change.

Marine cloud brightening: regional applications.

The general principle behind the marine cloud brightening (MCB) climate engineering technique is that seeding marine stratocumulus clouds with substantial concentrations of roughly monodisperse sub-micrometre-sized seawater particles might significantly enhance cloud albedo and longevity, thereby producing a cooling effect. This paper is concerned with preliminary studies of the possible beneficial application of MCB to three regional issues: (1) recovery of polar ice loss, (2) weakening of developing hurricanes and (3) elimination or reduction of coral bleaching. The primary focus is on Item 1. We focus discussion herein on advantages associated with engaging in limited-area seeding, regional effects rather than global; and the levels of seeding that may be required to address changing current and near-term conditions in the Arctic. We also mention the possibility that MCB might be capable of producing a localized cooling to help stabilize the West Antarctic Ice Sheet.

Marine cloud brightening.

The idea behind the marine cloud-brightening (MCB) geoengineering technique is that seeding marine stratocumulus clouds with copious quantities of roughly monodisperse sub-micrometre sea water particles might significantly enhance the cloud droplet number concentration, and thereby the cloud albedo and possibly longevity. This would produce a cooling, which general circulation model (GCM) computations suggest could-subject to satisfactory resolution of technical and scientific problems identified herein-have the capacity to balance global warming up to the carbon dioxide-doubling point. We describe herein an account of our recent research on a number of critical issues associated with MCB. This involves (i) GCM studies, which are our primary tools for evaluating globally the effectiveness of MCB, and assessing its climate impacts on rainfall amounts and distribution, and also polar sea-ice cover and thickness; (ii) high-resolution modelling of the effects of seeding on marine stratocumulus, which are required to understand the complex array of interacting processes involved in cloud brightening; (iii) microphysical modelling sensitivity studies, examining the influence of seeding amount, seed-particle salt-mass, air-mass characteristics, updraught speed and other parameters on cloud-albedo change; (iv) sea water spray-production techniques; (v) computational fluid dynamics studies of possible large-scale periodicities in Flettner rotors; and (vi) the planning of a three-stage limited-area field research experiment, with the primary objectives of technology testing and determining to what extent, if any, cloud albedo might be enhanced by seeding marine stratocumulus clouds on a spatial scale of around 100×100 km. We stress that there would be no justification for deployment of MCB unless it was clearly established that no significant adverse consequences would result. There would also need to be an international agreement firmly in favour of such action.