| 06 Oct 2025
Biochar-Induced Soil Property Changes May Reduce Temperature and Precipitation Extremes: Insights from Earth System Model Experiments
Abstract. Biochar has been proposed as a promising soil amendment for climate change mitigation, owing to its capacity to sequester carbon and alter soil physical properties. This study investigates the potential influence of biochar-induced changes in soil hydrological and thermal properties on future climate, with a focus on extreme climate events. We implemented a series of biochar addition scenarios (ranging from 5 to 150 t/ha) into the Max Planck Institute Earth System Model (MPI-ESM), modifying eight soil physical variables via pedo-transfer functions to investigate their impacts on climate in the near future (2040–2049) under the CMIP6 framework.
Our results show that while biochar-induced soil property changes produced minimal global effects on temperature and precipitation, they led to more structured and consistent responses in climate extremes over land. In particular, the addition of biochar reduced temperature extremes – especially nighttime minimum temperatures (TNn) – across cold regions such as Eastern Europe, the Russian Arctic, and West Siberia. Contrary to our initial hypothesis, these effects were not driven by enhanced latent heat flux but rather by increased humidity and cloud cover that altered surface energy balance via sensible heat redistribution. Precipitation extremes also responded to biochar addition, with a consistent decrease in extreme rainfall (Rx1day) over land. However, changes in consecutive dry days (CDD) were more region-specific, with increases observed in arid and coastal regions such as the Arabian Peninsula and Central Australia, indicating heightened drought risks in already vulnerable zones.
These findings suggest that although biochar’s direct modifications are localized, its indirect effects on climate extremes can extend across sensitive regions through land-atmosphere interactions. Our study highlights the importance of integrating both biogeophysical and biogeochemical pathways in Earth system models to better evaluate biochar's climate mitigation potential.
The article presents a sensitivity analysis to soil hydraulic and thermal properties carried out with the Earth System Model (MPI-ESM). These soil properties are modified to mimic effects of biochar addition over agricultural lands. The manuscript is generally well presented and the topic could be interesting – even in the case of “negative results”, i.e., biochar addition effects on climate are negligible; however, I have a number of comments that I think needs to be resolved before considering the manuscript further.
1) The effect of biochar is considered only in terms of increment in SOC and hydraulic and thermal properties are modified accordingly using PTFS. This is a limitation already mentioned by the authors (LL 379-382). As largely the manuscript is a sensitivity analysis to soil properties, I wonder if some other modifications induced by biochar as changes in pore size distribution or saturated hydraulic conductivity could be somehow imposed from literature, rather than passing through PTFs to represent a more realistic change of soil properties unrelated to SOC. For instance, in Table 1 saturated hydraulic conductivity somehow decreases in the various model scenarios while biochar addition literature suggests an increase.
2) In the numerical experiment biochar are added to all agricultural regions with various applications rate 5 to 150 ton/ha. Current biochar production capacity, according to my estimate (please double check), would be just sufficient for 700 km2 to 24 km2 of land at these application rates, vastly different than the size of global agricultural land. In other words, while the manuscript is a nice numerical experiment, it is very unlikely a realistic scenario, even in a distant future. This should be at the very least discussed.
3) Even more important, most of the reported differences have a relatively small magnitude. Now it depends exactly on how the experiment was implemented but, somehow by perturbing the soil properties, one perturbs the initial conditions of the ESM simulations, which means that the base simulations and various scenarios do not differ uniquely because of soil properties but also because of internal climate variability (two different stochastic trajectories of climate). For air temperature this might not be a big problem, but usually, internal climate variability can create differences (even over 40 year averages) for precipitation statistics and extremes in general that can be of the magnitude of reported changes or larger, thus I wonder how many of these changes are indeed a clear signature of soil properties rather than simply internal climate variability. Without a quantification of stochastic uncertainty, results as table 2 are not really useful. If the authors have imposed specific controls to avoid effects of internal climate variability, it needs to be explained, but I doubt this is the case. If changes in climate induced by soil properties are smaller than internal climate variability, this is still a result worth publishing in my view, but the narrative of the manuscript should change.
4) Finally, the mechanistic explanation provided by the authors, i.e., redistribution of sensible heat and modulation of atmospheric boundary layer dynamics, increases in cloud cover and surface humidity and reduction in outgoing longwave radiation are not entirely explained. First of all, if latent heat does not change, sensible heat changes do require a change in net radiation which is not fully explained or justified. Second, if latent heat is not changing what is causing the increase in near surface humidity. The authors need to dig deeper into the mechanisms causing the reported changes.
Minor comments
LL 77. Please note that 200 km is an extremely coarse resolution, and unlikely to be able to properly represent effects of soil properties on hydrological dynamics, which are likely to manifest more in models able to solve for lateral water redistribution and groundwater dynamics.
LL 85. Based on subsequent parts of the manuscript and especially caption of Fig. 3 where it is finally clear from the variable names, these should be water content at field capacity, water content at saturation, water content at wilting point. “Field saturation” by itself does not mean anything. It is also unclear why initial and maximum moisture content are treated as parameter and especially how maximum moisture content is different from “saturated water content”, which would be the correct name.
LL 89. I would suggest mentioning already which PTF is used.
LL 101.Which future emission scenario was used?
LL 113. What is the total area in km2 upon which biochar application was simulated?
LL 127-135. Please make clear, what are the inputs/outputs that are used in each PFTs. This part is confusing. Field saturation should be “saturated water content”, and field capacity, “water content at field capacity”; the same for water content at wilting point. Which water potential values were selected to determine water content at field capacity (-33kPa?) and water content at wilting point (-1500kPa?). This should be clarified. Saxton and Rawls pedotransfer functions do not compute a residual water content. The water potential value associated with the residual water content needs to be specified somehow. Please note that wilting point will depend on the specific vegetation type – not only on the soil, it would be better to define the residual water content, taking some very low water potential (i.e., -10MPa), but of course it depends on the detail of MPI-ESM implementation. It would be better to explain this part properly.
LL 135-136. This part is not well explained, how exactly these ratios were used?
LL 153. Water content is a prognostic variable – not a parameter or a property that can be derived from SoilGrids.
LL 205. Why Ksat is decreasing? More SOC should create more and larger pores and a more hydraulically conductive soil overall.
LL 250-260. The authors use throughout the manuscript variable names equal to the model definition of these variables, for instance, ahfs for sensible heat, I would suggest using in the figures and text the more common nomenclature and symbols, i.e., H for sensible heat, λE for latent heat, etc., it will make the reading more intuitive.
LL 340-342 and 360-362. It is not very clear what these percentages represent.
LL366. I am not sure SOC encompasses all the first order effects of biochar addition in changing the soil properties. I would rather argue that even if biochar addition will not modify SOC at all, soil hydraulic and thermal properties would still be considerably influenced, so not all the first order effects are captured.