the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Jan 2026
AIM-ALPHA v1.0: A partial equilibrium model of global agriculture and land use at basin-level resolution
Abstract. This paper introduces AIM-ALPHA v1.0, a global partial equilibrium model that explicitly represents agricultural markets and land-use dynamics with high spatial resolution at the national and subnational river-basin levels. The model covers 166 countries on the demand side and 400 production units on the production side, balancing supply, demand, prices, trade, and land allocation across 23 agricultural commodities from 2015 to 2100. AIM-ALPHA is a comprehensive framework that analyses global socioeconomic and environmental challenges such as food security, climate change, biodiversity loss, and their mitigation. This paper describes the model structure, evaluates performance against historical data, and demonstrates application via a case study on the impacts of climate-change mitigation on global food security. Validation shows that AIM-ALPHA reproduces the historical statistics around 2015 and remains within ±10 % of observed values up to 2020, confirming numerical consistency. The baseline projections are broadly consistent with those of other integrated assessment and agricultural economic models. We also present a scenario analysis using a 1.5 °C climate-change mitigation pathway, which incorporates global carbon pricing, bioenergy expansion, and afforestation. The results show that agricultural production declines while food prices rise, reducing calorie intake, particularly in developing regions. The country-level resolution of the model reveals considerable heterogeneity within aggregated regions, with several developing countries in Africa and Asia facing particularly large reductions in food availability. Such distributional effects are largely obscured in more regionally aggregated models. These findings highlight the trade-offs between climate-change mitigation and food security and the importance of national-level assessments that capture socioeconomic and spatial heterogeneity. AIM-ALPHA serves as a comprehensive platform when exploring linkages between human and environmental systems and analysing sustainable land-use transitions and policy design at national and regional levels.
- Preprint
(3018 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-5911', Anonymous Referee #1, 13 Feb 2026
- AC1: 'Reply on RC1', Ryo Totake, 24 May 2026
-
RC2: 'Comment on egusphere-2025-5911', Anonymous Referee #2, 18 Mar 2026
This paper presents an ambitious effort to develop a new partial-equilibrium model for agriculture and land use to be integrated in the AIM integrated assessment model framework. The comprehensive development is impressive and the paper and model documentation are a good basis for follow-up work. Some components of the model are unclear though and need further explanation, and some model dynamics are counter intuitive. Generally I do commend this work and support publication after my concerns have been addressed.
General comments:
- The resolution of the model is at the country and basin level. It is unclear however which dynamics are represented at the country-level and which at the basin-level and how these dimensions interact. Also the benefit of the basin-level at this stage of the model is unclear as water does not seem to be represented. Please explain better how this works and why this choice is made.
- Allocation is driven by land rents (section 2.7) which is therefore quite a fundamental variable. What is this information based on? Please explain better how this works.
- The validation to 5 years of data is underwhelming. It would be very valuable if a back casting exercise could be added, e.g. from 2015 to 2000 to assess model dynamics based on a longer time series.
- Regarding the results, notably the pasture trend is very surprising. Historically pasture has been gradually declining for quite some time, but this model projects a steady increase from 2015-2040 in all scenarios. In line 360 it is stated that this depends on the ratio of roughage to concentrate feed which in turn is based on GDP/capita, but then this relationship does not seem to reflect recent trends as GDP has increased while pasture has decreased… I would say this really needs to be improved as it dominates the land use results. Similarly on the regional level the cropland area trend from recent trends in LAM and MEA. How can this be explained and can it be improved?
Specific comments:
- Line 110: figure 1 shows water and nitrogen resources but these are not mentioned at all in the methodology section. How are they represented?
- Line 144: I am a bit puzzle by the processed foods category ‘sweeteners’. This also includes sugar right, so it would be better to call it ‘Sugar and sweeteners’ as sugar is the bulk of the product. Also it now sounds like inputs for sweeteners can only come from sugar crops (‘primarily’), while maize is also in important input (10-12%), especially in the USA. Is this accounted for? If not, why is this a reasonable assumption? If it is accounted for this should be better described.
- Line 236: also here the assumption on input for processed foods is unclear. Is it calibrated to historical shares? Or assumed to be 1 primary crop only?
- Line 261-262: can you give more details on the available land constraints. Are there more than the two constraints mentioned? How much does this restrict the model? Are there also minimal yields constraints or is that taken into account through prices? And how are protected areas taken into account?
- Line 266: it is unclear to me what calculations happen at the country-level and which at the river basin level. I assume trade is at the country-level, but how is this then connected to the basins? This needs to be made more clear throughout the methodology.
- Line 300: ‘SSP1 models’ is a weird phrasing, reformulate to ‘SSP1 represents’ or similar.
- Line 307: this section mostly describes the SSP narratives which better fits in section 3. It would be useful to describe takes population and gdp changes into account.
- Line 379-380: it is stated that ‘livestock consumption is higher because of the assumptions in table 1’. But table 1 does not explain this in my view as only the gdp-based food demand relationship is mentioned. The difference in SSP1 is most likely caused by omitting the assumption of reduced preference for livestock products that is typically adopted.
- Line 495-497: This statement is correct, as figure 8 show for quite a few variables that the results are outside the SSP database range.
Textual comments:
- Line 55: replace ‘capture of’ by ‘representation of’ or something similar.
Citation: https://doi.org/10.5194/egusphere-2025-5911-RC2 - AC2: 'Reply on RC2', Ryo Totake, 24 May 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,183 | 396 | 122 | 1,701 | 114 | 136 |
- HTML: 1,183
- PDF: 396
- XML: 122
- Total: 1,701
- BibTeX: 114
- EndNote: 136
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General Evaluation
This manuscript presents AIM-ALPHA v1.0, a global partial equilibrium model designed to simulate agricultural production, land use, trade, and prices, with explicit representation at national and river-basin levels. The model integrates supply, demand, and land allocation decisions across multiple commodities and regions, and is positioned as a tool for analyzing long-term interactions among agriculture, climate mitigation, and food security. The paper provides an overview of the model structure, core assumptions, and computational framework, and evaluates model performance against historical statistics. The authors demonstrate the model’s application through a climate mitigation scenario and emphasize that fine spatial resolution allows the identification of distributional effects that are masked in more aggregated models. The use of a GAMS/MCP formulation and the provision of model code are also notable strengths, as they support transparency and reproducibility in principle.
Overall, the manuscript introduces a relevant and timely modeling framework, and the development of a basin-level, globally consistent partial equilibrium model represents a potentially valuable contribution to the literature. The model has clear potential applications in food security analysis, land-use policy assessment, and integrated climate-economy studies. The paper is generally well organized and provides substantial technical material. However, despite the ambitious scope and policy relevance of the framework, several aspects of the documentation, validation, and methodological transparency require further clarification and strengthening. In particular, more rigorous evaluation of model performance, clearer exposition of the underlying economic structure, and a deeper discussion of uncertainties and limitations are needed to enhance the credibility and usability of the model. Addressing these issues would substantially improve the manuscript’s contribution to the GMD community.
Major Comments
(1) Positioning, Contribution, and Literature Coverage
The manuscript does not sufficiently situate AIM-ALPHA within the existing literature on spatially explicit partial equilibrium and land-use models. While MAgPIE and GLOBIOM are mentioned, other closely related modeling frameworks, such as GTAP-AEZ-RB, SIMPLE-G, and GTAP-SIMPLEG, are not discussed. Methodologically, AIM-ALPHA appears to share substantial similarities with the GTAP-AEZ and SIMPLE modeling traditions, particularly in its treatment of land heterogeneity, spatial disaggregation, and market equilibrium. In addition, the authors rely on USDA-based crop elasticities that are widely used in these existing models. Given these overlaps, the manuscript would benefit from a more systematic comparison with these frameworks. A more comprehensive literature review and explicit discussion of similarities, differences, and added value are needed to clarify the model’s contribution.
(2) Economic Representation
Figure 3 presents price on the horizontal axis and quantity on the vertical axis, which is inconsistent with standard economic conventions, where price is typically shown on the vertical axis. While this may seem minor, it raises concerns about the authors’ familiarity with standard economic representations and modeling practices. The authors should ensure that the economic interpretation of the diagram is correct and consistent with the broader literature. More generally, the paper would benefit from stronger engagement with established economic modeling conventions.
(3) Model Transparency, Documentation, and Structural Assessment
The manuscript provides extensive code and scripts, but very limited formal documentation of the model structure. There are essentially no explicit equations describing behavioral relationships, market-clearing conditions, or closure rules. As a result, it is currently impossible for reviewers and readers to evaluate the internal logic and structural consistency of the model. Without a clear mathematical representation, one cannot assess whether the implementation faithfully reflects the intended economic structure. For a GMD paper, this is a significant limitation. The authors should provide a formal description of the core model equations, including supply and demand systems, land allocation, trade representation, and equilibrium conditions. This is necessary to allow independent assessment and reproducibility. In future publications, you can refer to this paper for equations.
(4) Baseline Replication and Validation
The main evidence for model validation appears to be Figure 10, which compares estimated values with historical FAOSTAT data. However, this comparison raises important concerns. The base year of the model is 2015. In standard partial equilibrium models implemented in GAMS using MCP formulations, the benchmark equilibrium should “exactly” replicate the base-year data when no shocks are applied. This is a fundamental property of calibrated equilibrium models. Figure 10 indicates substantial deviations between modeled and observed values even in the base year, with many observations falling outside narrow tolerance ranges. This suggests that the model does not properly replicate its calibration data. In many established modeling frameworks (e.g., GEMPACK, GAMS/MCP and MPSGE-based models), baseline consistency is routinely verified through numeraire checks and zero-shock simulations, which reproduce the reference dataset exactly. Deviations at the benchmark stage often indicate structural inconsistencies, coding errors, or incomplete calibration.
The authors should therefore: Explain why exact base-year replication is not achieved; Clarify whether the reported discrepancies arise from deliberate modeling choices or technical limitations; and provide additional diagnostic checks demonstrating numerical consistency. Without convincing evidence of proper benchmark replication, the credibility of scenario results is difficult to assess.