| 13 Jan 2026
This is FRIDA
Abstract. FRIDA is a new contribution to the portfolio of integrated assessment models (IAMs) that address the climate – energy – economy – and society nexus. The FRIDA acronym stands for Feedback-based knowledge Repository for IntegrateD Assessments. By naming it a "knowledge repository" we signal that the FRIDA model is never finished; it represents the current state of knowledge of the development team at any given time. We aim to continually integrate new scientific findings to keep the FRIDA up to date.
FRIDA comes with a learning environment that, together with the model's low computational cost, makes it a useful tool for education. It can be used in the classroom setting in interdisciplinary climate science courses and will allow students to understand how their discipline is intricately woven into the rest of the climate science disciplines. This feature set makes FRIDA accessible to a wider range of users than just researchers and scientists. Our aim is to lower the barrier to entry of using this model so that even lay people are able to use the model to build an understanding of the interconnectedness of climate and humans. Additionally, the low computational burden allows for uncertainty exploration by varying model parameters.
In this collection of papers in the Geoscientific Model Development (GMD) journal we intend to document the developments of FRIDA, from its origin in the years 2023–2026 within the European Horizon project "WorldTrans – Transparent Assessments for Real People" (FRIDA version 2.1 and FRIDA V3); and (hopefully) future versions that the spirited (and growing) development team will hopefully ensure. The intention of this brief introductory paper is to provide the contextual framework for the original model, and to explicitly state the original requirements.
For this paper as a stand-alone paper, I don't have much to recommend for revisions; as noted in the section 1 (Introduction), this paper is intended itself as a preliminary introduction to a number of papers, some of which are already written and available online. I did not review those other papers in support of providing these comments, so I don't know if any changes are warranted, to address my questions below.
The paper is written as if by a team ("we"), but there is only one author listed. Perhaps it should be clarified up-front that this author is writing on behalf of a larger team? The language is at times informal, not consistent with how scientific papers are normally written ("Do remember, though,..."). There's an uncapitalized sentence fragment in line 56, and an underlined sentence in lines 59-60. Lines 51-54 have two back-to-back single-sentence paragraphs.
My biggest questions pertain to the credibility of the projections coming from the model. However, this specific introductory paper is probably not the place to address this concern, and to provide the comparisons between FRIDA and the existing body of projections to 2100. Similarly I am wondering about the key data inputs to the model. Is it trained purely on statistical and historical data? How many of the equations used for projections were pulled from the literature or other models, as opposed to developed by this team? Are the parameterization designed with reference to existing projections from other models?
Within the paper, the only results shown are in Figure 7, where the figure is too small/blurry to interpret; the purpose of the figure seems primarily to show that results can be produced. Still, for a paper documenting the existence of a new model, it seems pretty basic that some projections of common scenarios ought to be run and compared with the existing published body of scenarios (e.g., AR6). Again, if this is done in subsequent papers then there is no issue here.
On a similar note, the author continually stresses the very high portion of endogenous variables in the model, i.e. that there are very few exogenous variable specified. This seems a worthy aspiration in general, but care should be taken to ensure that making a variable endogenous actually improves the quality of the projected outcomes. In the current generation of IA models, the major exogenous assumptions (e.g., population, GDP, techno-economic characteristics of energy technologies) can act as guardrails on future modeled outcomes. Some acknowledgment of this fact seems warranted, along with discussion of whether constraints on variables were applied, whether exogenously or simply within the equations used for projecting.