| 11 Jun 2025
Phenology, fluxes and their drivers in major Indian agroecosystems: A modeling study using the Community Land Model (CLM5)
Abstract. Agroecosystems are the largest land use category, covering more than half of the land surface in India, yet the understanding of spatio-temporal variability of the terrestrial fluxes over these ecosystems is limited. Previous studies are mostly at the site scale, relying on eddy covariance observations that fail to capture the spatial variations across diverse climatic regions of India. The only regional-scale study, Reddy et al. (2023), is limited to wheat crops and lacks the robust model calibration, leading to higher uncertainties in simulated crop physiology and carbon uptake across diverse climatic regions. This study is the first to comprehensively investigate long-term trends (1970–2014) in crop physiological parameters and terrestrial fluxes across major croplands of India. This study uses a robustly calibrated Community Land Model version 5 (CLM5) to conduct numerical experiments for understanding the influence of natural and management factors on crop physiology and terrestrial fluxes. CLM5 simulations show Pearson's correlation coefficients exceeding 0.6 for regional carbon fluxes and 0.95 for regional yield estimates. The results show that crop physiology parameters have increased more than twofold since the 1970s, with crop carbon uptake by agroecosystems doubling, while respiratory losses decreased due to improved nitrogen fertilization. The largest impact is due to nitrogen fertilizer usage and nitrogen-related processes, which contributed to more than 50 % of the observed trend in crop physiology parameters and carbon uptake in both rice and wheat. Followed by irrigation application and increasing atmospheric carbon concentration. The results further reveal that CLM5 performs particularly well in estimating carbon fluxes during the cold, dry rabi season and simulates water and energy fluxes more accurately during the warm, wet kharif season. The results highlight the need to investigate the stomatal activity for crops in CLM5 and understand the reason for comparatively poor simulation of carbon fluxes in the kharif season and water and energy fluxes in the rabi season. This is the first study to address both the spatial and temporal variations in agroecosystem physiology and fluxes in India using a robustly calibrated and evaluated land model. Given the scarcity of studies on terrestrial fluxes in tropical agroecosystems, this work demonstrates the importance of using limited site-scale data to improve regional-scale models and enhance our understanding of tropical agroecosystems.
The manuscript entitled ‘Phenology, fluxes and their drivers in major Indian agroecosystems: A modeling study using the Community Land Model (CLM5)’ benchmarked model simulated carbon, water, and energy fluxes across croplands in India at both site and regional level; following model validation, the authors examined impacts of climate change, elevated CO2, nitrogen fertilization, and irrigation on the long-term trends (1970-2014) of matter and energy fluxes across Indian croplands by conducting four model simulations. The authors concluded that N fertilization and irrigation are top drivers for the increasing trends of carbon fluxes with additional effects from elevated CO2. This work has potential management implications for agricultural production and food security under changing climate. The work is straightforward but needs some clarifications in modeling methods; the results section can also be improved to be more concise. See my comments below.
There are a lot of figure caption-style of writing when documenting results (e.g., L274-275, L301, L330 etc.). These should be moved to the corresponding figures as caption instead of being documented as results.
When documenting the trends and drivers on simulated trends (e.g., section 3.2 and 3.3), most of the patterns in carbon fluxes and crop physiology are consistent. So I think it can be more concise to combine results instead of documenting them separately. For instance, L390-430 can be combined into one paragraph by documenting the trends of GPP, AR, and NPP, and then how climate, CO2, N fertilization, and irrigation influence them correspondently. For instance, CO2 has significant positive influences on all carbon fluxes with effect size being 9, 3, 6 gC/m2/year for GPP, AR, and NPP respectively.
In several places the authors have stated how well CLM can capture the seasonality of cropland function (e.g., L314, L618, L621), but seasonal pattern is not shown in any of the figures included in main. Maybe there can be a better way to visualize the results if that’s something authors would like to highlight.
Some other comments:
8.L488: not wheat but rice?