Reversing soil degradation by 10% could feed 70 million more, global analysis finds Sadie Harley Scientific Editor Robert Egan Senior Editor Factors such as erosion, dehydration and soil compaction are jeopardizing food supplies for broad swaths of the world's population, according to a study by the University of Bonn, the University of Minnesota, the University of Basel and Hamburg University of Technology. If human-made soil degradation were reversed by 10%, Earth's arable land could feed an extra 70 million or so people in purely mathematical terms. The findings have been published in the journal Nature Food.

Not only are slash-and-burn agriculture, overfertilization and large-scale monoculture threatening ecosystem biodiversity, but many current farming practices are also doing long-term damage to soil quality. This problem is being exacerbated by human-made crises such as climate change, with storms and heavy rainfall causing fertile topsoil to be blown or washed away. In addition, groundwater levels are falling, not least because precipitation often comes in such heavy concentrations that it simply runs off instead of seeping into the ground.

Degraded soil is known to be less fertile. However, the severity of this effect and its impact on crop yields have been something of a bone of contention until now. "Now, for the first time, we've attempted to quantify the consequences of degradation for all the arable land around the world," explains Dr.

Hadi Hadi, a postdoctoral researcher in the Land Economics Group at the University of Bonn, who carried out many of the analyses for the recently published study. How much could unaffected soils yield? The researchers involved in the study started by asking themselves how high yields could be in practice without any degradation.

To this end, they first divided all the arable land on the planet into squares 10 kilometers (6 miles) long and wide. "We then looked for squares with similar prevailing conditions, i.e., comparable precipitation levels and temperatures, and compared the yields in these areas," Hadi says. Adjusting their figures to account for differences in the use of fertilizer, irrigation and other cultivation factors enabled the researchers to identify areas with particularly high yields.

Many of these were also those that had sustained hardly any soil damage up to that point. "Thus, these areas give us a benchmark for what's possible in soils that haven't been degraded or influenced in any way," explains Professor David Wuepper, head of the Land Economics Group. Machine learning uses yield gaps to calculate degradation rates For each 10 × 10 km parcel of land, the working groups involved were able to calculate a specific "yield gap," which they then correlated with five different parameters connected with soil degradation: erosion, dehydration, compaction caused by the widespread use of heavy agricultural machinery, the loss of carbon dioxide in the soil and minimal vegetation cover.

"Rather than gathering this data ourselves, we drew on detailed studies by various working groups," Wuepper points out. With the aid of complex AI-powered evaluation methods, the researchers were able to use the yield gap they had identified to calculate the degree of soil degradation in each parcel.