| 12 Aug 2026
Sea level rise in Los Santos and Los Angeles: a comparative flood analysis using Grand Theft Auto V
Abstract. We present the first quantitative GIS-based flood inundation analysis using terrain data extracted from a commercial video game, applied to the fictional city of Los Santos from Grand Theft Auto V (GTA V) under two contrasting IPCC Fifth Assessment Report (AR5) Representative Concentration Pathway (RCP) scenarios (RCP 2.6, strong mitigation, and RCP 8.5, high-emission baseline) combined with historical storm surge records from the Los Angeles tide gauge, across three time horizons (2100, 2300, and 2500). Los Santos serves as an ideal case study because it faithfully replicates Los Angeles (LA), enabling direct comparison of simulated flood outcomes in the game world with published scientific vulnerability analyses for LA. Results indicate that Los Santos is modelled with sufficient geomorphological detail for projected flooding patterns to closely resemble those identified for Los Angeles, including the spatial sequence of inundation across the marina district, the Pacific Palisades coastline, and the port. Under RCP 8.5 conditions with storm surges, combined sea levels of up to 5.18 m by 2300 would inundate 5.1 % of the mapped land area; in a longer-term illustrative projection to 2500, extreme sea levels of 8.22 m would inundate 6.5 %, rendering the port, wastewater infrastructure, and coastal energy facilities permanently non-functional. The mitigation scenario RCP 2.6 limits inundation to 2.2 % of land area at 2300 with no major infrastructure at risk. We evaluate the potential of this approach for geoscience communication and education, arguing that the intrinsic familiarity of the Los Santos environment provides an affective engagement with climate change impacts that is difficult to achieve through conventional visualisation alone.
Thank you for the opportunity to review this paper. It falls right in my area of interests and specialisms, games and flood modelling, and I enjoyed reading it very much. I am recommending that the paper is accepted for publications by Geoscience Communication after minor revisions.
In this paper the author’s use a community-derived recreation of the terrain height map used for the popular Grand Theft Auto V game (GTAV). The game world, the fictional city of Los Santos, is heavily inspired by the real world Los Angeles, sharing similar features that might by familiar to players. A simple bathtub model is applied to the Los Santos heightmap and flood simulations compared to those projected for Los Angeles, with the authors showing that similar patterns emerge in both. The authors then argue this has potential educational benefits based on familiarity from players and psychological closeness with climate change.
The paper is successful in demonstrating that the game’s heightmap is sufficiently detailed and constructed to allow reasonable predictions of future sea levels possible. This will have interest to educators and communicators in geosciences who are considering using game-based approaches as a tool, and also to games developers who may wish to use realistic climate projections and environmental change as a plot point or game challenge.
From my own experience of working with game developers, it is likely that the GTAV devs will have incorporated freely available elevation data in the development of the terrain mesh for example SRTM. This would strengthen your arguments but without speaking to the developers we cannot know.
There are a few issues that the authors will need to address before it can be published in Geoscience Communication: the proposed educational application; the language around simulations; and need for wider critique of bath tub models. My recommendations are:
My recommendation is the authors replace 5.3 with a shorter (one or two paragraphs) discussion on possible (and appropriate) applications of the game in education and outreach. I suggest the authors consider non-formal routes, for example, the projections might be of interest to game-based and climate creators, including Any Austin or Simon Clark.
“Under RCP 8.5 conditions with storm surges, combined sea levels of up to 5.18 m by 2300 would inundate 5.1 % of the mapped land area; in a longer-term illustrative projection to 2500, extreme sea levels of 8.22 m would inundate 6.5 %, rendering the port, wastewater infrastructure, and coastal energy facilities permanently non-functional.” Lines 20-23.
This is a statement in model land because it is presented as a reality with no possibility that the model might be wrong. The impacts presented are not certain and should not be presented as such. Instead, it should instead be something like:
“Under RCP8.5 scenario with storm surges, combined sea levels are projected to be 5.18 m by 2300, with the model showing 5.1% of the mapped land area inundated; for a 2500 scenario with projected sea levels of 8.22 m, this increases to 6.5 % modelled inundation, which places the port, wastewater infrastructure, and costal energy facilities at risk.”
This shift in language should be made throughout the manuscript.
Thompson, E., 2022. Escape from Model Land: How Mathematical Models Can Lead Us Astray and What We Can Do about It. Basic Books, London.
Ramirez, J. A., Lichter, M., Coulthard, T. J., and Skinner, C. J., 2016. Hyper-resolution mapping of regional storm surge and tide flooding: comparison of static and dynamic models. Nat Hazards 82, 571–590. https://doi.org/10.1007/s11069-016-2198-z
Sanders, B. F., Wing, O. E. J., & Bates, P. D., 2024. Flooding is not like filling a bath. Earth's Future, 12, e2024EF005164. https://doi.org/10.1029/2024EF005164
For additional context, not necessarily for inclusion in the manuscript, there has been heavy criticism for bath tub (or similar) model derived flood maps as part of public engagement with climate change, with them being a potential source of misinformation if used inappropriately. For example, from Prof Ivan Haigh (https://tyndall.ac.uk/news/will-london-soon-be-underwater/) and myself (https://geoskinner.games/2017/05/11/no-easy-answers-below-sea-level-does-not-mean-below-the-sea/). This should be a consideration for any educational application of this work.
Line-by-line corrections:
Line 46: Geo-games and geo-gaming already have a specific meaning, being games are based on a player’s current location (e.g. Pokemon Go). I’d avoid using the term here. I’d suggest game-based approaches in geoscience instead (not very catchy…).
Line 70: The heightmap was not extracted from the game it was reproduced. This should be clear here.
Table 1 (and elsewhere): State the indicative present-day magnitude of the storm surge applied here.
Figure 4: Consider the accessibility of the shading used here. Green and red is not the best combination.
Line 331: You are speculating about an editorial compression of the landscape – you would need to speak to the devs to know this. Make it clear this is your interpretation of what has happened (I don’t disagree btw).
I look forward to seeing the revised paper and the final paper in publication.
Chris Skinner