Can we treat hurricanes like tumors?
What happens when precision-medicine playbooks rewrite parametric insurance?
👋 Welcome to Climate Drift: your cheat-sheet to climate. Each edition breaks down real solutions, hard numbers, and career moves for operators, founders, and investors who want impact. For more: Community | Accelerator | Open Climate Firesides | Deep Dives
Hey there! 👋
Skander here.
Picture this: you can calculate the impact of a Category-4 hurricane before the clouds even swirl, or tailor a cancer treatment the moment a rogue gene flickers on the screen. Different arenas, same game-changer: data that shrinks uncertainty to a single, decisive moment.
After fifteen years at the leading edge of precision medicine, decoding brain circuitry, shepherding AI diagnostics through the FDA, and translating biomarkers into bedside choices, driftie Nirmal Keshava is returning to his original orbit: Earth.
This time he’s trading MRIs for multispectral satellites and dosing schedules for parametric triggers, but the mission stays the same: convert cutting-edge science into faster, fairer decisions that safeguard lives and livelihoods.
In this breakdown, he shows how to:
🛰️ Track Hazards Like Tumors
From hurricane wind speeds to rogue DNA, why better sensors beat blunt averages every time.
⚖️ Make Claims Before the Clouds Burst
Parametric triggers = instant payouts. Think Venmo for catastrophes.
🧬 Borrow the Biomarker Playbook
Composite climate indices that slash basis risk just like multigene panels revolutionized oncology.
🤖 Plug AI Into Both Sides
Random forests, deep nets, FDA approvals, now repurposed to predict floodplains and wildfire corridors.
💰 Align Incentives, Unlock Capital
Insurance underwriting as the hidden lever to bankroll adaptation at planetary scale.
If you want to understand climate chaos with the precision of a cancer diagnosis, this is your field guide.
🌊 Let’s dive in
But first: Who is Nirmal?
Nirmal has built and led multi-disciplinary data science teams across industry, academia, and government, with a track record spanning defense, security, healthcare, pharma, neuroscience, mental health, and telecom. The work has led to numerous publications and patents, and has been recognized by the National Academy of Engineering, the National Science Foundation, and IEEE.
After 15 years at the forefront of life sciences, working in neuroscience, oncology, and precision medicine, he is now returning to climate-focused challenges. “Returning” because his early career began with AI-driven satellite imagery at CMU (in collaboration with NASA) and at MIT Lincoln Laboratory, long before the commercial geospatial climate industry existed.
Recently, he has re-engaged with the climate space by completing the AirMiners and Climate Drift accelerators, joining the Engineering for One Planet coalition, and becoming a facilitator for the En-ROADS climate solutions simulator. His focus now is on bringing a systems-level, strategic mindset to climate problems at the intersection of AI, economics, innovation, and international development.
Can we treat hurricanes like tumors?
This is a long one again, click the title if you want to read it fully 👆
“To me, boxing is like a ballet, except there's no music, no choreography and the dancers hit each other.” Jack Handey
It might seem like a stretch to think that climate risk modeling for parametric insurance and precision medicine are comparable, but they share similar trajectories and motivations. They work at opposite scales, one buried deep in our DNA and the other from low Earth orbit. Having worked at both, though, is an unique vantage point, one I hope can help the climate stakeholder community learn how health care has translated new science and data modalities into clinical care, particularly using AI.
Further, this survey highlights new model-based climate frameworks that capture the full continuum between core science and economic outputs. They paint a future where the impact of climate mitigation techniques can be modeled a priori to maximize the likelihood of success.
Finally, AI algorithms and architectures that are powering new FDA-approved clinical products are entering practice just as AI frameworks for specific climate catastrophes are emerging. Invariably, both health and climate are deeply scientific subjects, but data science is the key ingredient for achieving the accuracy and precision that makes plausible business propositions for both.
In this survey, we discuss the following:
Different Purposes, Common Goals: Precision medicine and climate risk modeling are harnessing scientific and computational models to replace a one-size-fits-all approach with efficient and precise frameworks to reduce human suffering.
Parametric Insurance is a Lever for Addressing Climate Change: The measurement of risk to create parametric insurance contracts is at the intersection of science, business, and AI/data science and can move markets towards greater sustainability.
Both Roads Go Thru AI/Data Science: Increasing use of different types of data, platforms, including remote sensing imagery, and AI algorithms can increase the precision of risk estimates and accelerate climate mitigation efforts.
Precision Public Health: The combination of precision medicine for climate effect mitigation are the foundations of a nascent new field.
Aligning Incentives for Climate Mitigation
The fear of an adverse diagnosis motivates us to make lifestyle and dietary changes to improve our prognosis. And, to insulate ourselves from the costs of treatments and hospitalization we sign up for health insurance. When the incentives align – our own and the institutions we depend on to maintain our health – there are better health outcomes individually and societally.
Market-driven approaches to climate risk and adaptation seek a similar alignment but are trickier because the outcomes we seek, e.g., a reduction in the rate of global temperature increase, better air quality, require global alignment and contributions. The Paris Climate Accords signed in 2016 did just that, securing voluntary targets from 196 countries to reduce their carbon emissions. Each country then can introduce the incentives that enable their targets to be met.
The question then becomes, between the extremes of a government mandate and full voluntary compliance, how can market-based incentives be created for households and stakeholder organizations that then collectively move countries toward their emission targets?
Health Care and Climate: On Parallel Journeys
Health care and climate are in the news daily and have a substantial impact on the US economy. As depicted in Figure 1(a), health care is comprised of the hospital and health care delivery networks, pharmaceutical R&D, medical devices, and several other associated industries. As of 2023, health care expenses were 16.5 % ($4.9 trillion) of the US GDP [1]. In 2021, 8.6% of people in the USA were uninsured, but that was distributed disproportionately across states and ethnic groups with 18% of Texas residents uninsured and with minority subpopulations all above the national average of uninsured [2].
In 2024, there were 27 billion-dollar (or more) climate-related disasters in the US, costing $182.7 billion [3]. Figure 1(b) illustrates the diversity of disaster types and geographic distribution, and there is evidence that climate effects will disproportionately impact minority and low-income individuals [4]. The most devastating event was Hurricane Helene, which incurred $79.6 billion in losses, yet, Helene’s destructive wind speeds of up to 225 km/h were not the main cause of the high claims burden [5]. In the storm’s wake, severe flooding from heavy rain unexpectedly spread northward into the Appalachian regions from Georgia to North Carolina. Statistically, the three days of extreme rainfall in North Carolina and Georgia would only occur on average once every 1000 years.
Two severe thunderstorm fronts that struck the Midwest in March and Texas in May, accompanied by numerous tornadoes, were among the costliest insured loss events of the year. Together they caused total losses of almost US$ 13bn, with around US$ 10bn of that insured. The figures confirm the trend: non-peak events such as severe thunderstorms are now causing cumulative damage equivalent to a severe hurricane year after year – with insurers bearing a significant share of the costs [5].
Both health and climate are capitalizing on the latest science and engineering to realize and deploy solutions to their respective challenges. We look more closely at important inflection points in each domain.
Turning a Corner in Medicine: Precision Medicine
The year 1998 was a big year for developing a new generation of drugs. Herceptin was approved by the FDA as the first targeted therapy, specifically engaging with the HER2 receptor which is implicated in many different cancers, including aggressive forms of breast cancer [6]. It heralded the arrival of precision medicine, which aims to overturn the existing one-size-fits-all approach to medicine by developing and delivering the right drug to the right patient at the right time.
A main driver for pursuing more precision is increasing evidence that many of the most frequently prescribed drugs do not help patients [7]. Figure 2 illustrates this for ten commonly prescribed drugs. Generally, this result shouldn’t surprise anyone since clinical trials are commonly run on a highly homogeneous and “clean” trial population under strict constraints; yet after FDA approval, they are prescribed to a far more diverse patient population. Succeeding initially with a narrow population isn’t a guarantee for broader success, and often that isn’t revealed until large-scale, retrospective epidemiological analyses are performed long after the drug has been approved.
In 2015, President Obama further underlined its importance with the Precision Medicine Initiative to connect genetics with disease and produce treatments that target specific genetic aberrations, frequently called “biomarkers [8].” One of the by-products of the Obama initiative was the All of Us research program which has collected longitudinal health data from over 1M volunteers on many scales (e.g., genetic, demographic, etc.). The hope is that advanced analytical methods, including causal modeling and AI/ML, could uncover new hypotheses in these deep and wide data sets about how diseases originate and proliferate in specific human subpopulations.
Climate Risk: Many Milestones, Maybe One Tipping Point
There have been important milestones in climate modeling, but the most impactful moments have been when climate science intersects with politics. The awarding of the 2007 Nobel Peace Prize to the Intergovernmental Panel on Climate Change (IPCC) for the series of four reports issued since 1990 and to Vice-President Al Gore for his tireless advocacy was a major step change in the recognition of climate change’s global importance. Instead of receiving one of the other scientific Nobel Prizes, their Peace Prize announcement specifically highlights the IPCC’s conclusions that “Climate Change will Increase the Danger of War.” In effect, the announcement states that climate change is not a scientific abstraction; it’s a human dilemma of either action or inaction.
Besides the international attention, the Nobel Prize validated many decades of scientific contributions to build ever more complex models operating on multiple scales. They inform many practical aspects of our daily life, especially weather, transportation, and agriculture, and are evolving to improve spatial resolution and prediction accuracy. Many multi-stakeholder ventures are aggregating climate and infrastructure into data sets for archival and systematic consumption [9] [10]. Recognizing the interconnectedness of air, ocean, and land and the processes that guide them has accelerated the development of deeply integrative models to describe natural climate processes on a global scale.
The Common Thread: Insurance and Risk
What ties health and climate together are how they both dovetail into a common global business proposition: insurance. At its core, insurance is a contract between two parties that transfers the risk of financial loss away from a person or business entity to another entity, the insurer. To be viable, the insurer must effectively assess specific risks and match premiums it collects with payouts it makes when losses occur due to insurable events within the bounds outlined in the contract. In everyday life, insurance insulates the owner of the policy from the cost of catastrophic health problems and catastrophic natural events that threaten our lives and property. Insurance contracts work when incoming premiums for an event with a certain likelihood are balanced by outgoing claim payments.
When causal linkages between genetics and disease are discovered, we improve our assessment of the risk of developing the associated disease, which otherwise would be at worst, completely unknown, and at best, empirical or anecdotal. Many examples exist of genes that indicate an elevated, hereditary risk of progressing to a disease, e.g. the BRCA, BRCA2 genes for breast cancer, and CFTR for cystic fibrosis [11]. So, a priori knowledge can raise ethical questions of how much insurers can use this information to deny insurance for circumstances that are wholly out of the control of patients; consequently, federal and state level regulations restrict how much genetic information can be used to develop premiums and deny coverage.
Property insurance creates a contract between underwriters and insurees based on predictive models of natural catastrophes, and the contract is most effective when risk estimates are both accurate and specific to each insuree, e.g., homeowners are paying premiums based on their actual risk instead of being lumped into a broader pool of homeowners with a wider range of risks. The best examples of “Nat Cat” models are for flood and fire which break down risk into four parts: hazard, vulnerability, exposure, and loss [12].
One important distinction between climate and health risk is that as climate patterns change, existing catastrophe models become increasingly mismatched [13] [14]. Outdated risk models can underestimate the frequency and severity of current climate disasters and the associated claims; consequently, property and casualty insurers are seeking new methods to accurately estimate risk as the damage from climate change events increases. This is precisely where next-generation natural catastrophe insurance models become data science problems.
A good example is agriculture. Recent work suggests that compound effects of climate change in the form of extreme heat and rain will increase and require substantial adaptation [15]. Figure 3 illustrates the compound influences on crop health. The conclusions suggest that compound stresses induced by climate change will introduce new modeling challenges requiring greater integration of different climate stressors and phenomena.
Barometers of Risk: Biomarkers and Climate Indices
Everyone who has had a fever knows the importance of the number 98.6 as a surrogate for good health and illness; in lieu of a thermometer, we place a hand on a forehead and often, that is good enough to know how to react. Similarly, climate scientists invoke the Keeling curve (see Figure 4) to convey increased levels of atmospheric CO2 as an index of climate health that, when it rises, is accompanied by increasingly dire consequences. A key construct for enabling precision medicine is to move from subjective assessments to quantitative measures to guide clinical decision making. Biomarkers are a measurable indicator of a biological state or condition, and they enable definitive “signposts” to guide a treatment decision in one direction or another. The best example is a genetic mutation, which can point to using a specific therapy for a disease, but more importantly, could prognosticate risk for developing a disease. The best biomarkers are those that are easily attainable and are a direct surrogate for how well a drug is performing or for the severity of a disease.
Keen to capitalize on their predictive capability, the insurance industry is already using biomarkers to make critical care and coverage decisions. For example, a specific type of biomarker testing for cancer subtypes (i.e., “tumor profiling”) is becoming particularly important because many newer cancer therapies target specific mutations (e.g., EGFR in non–small cell lung cancer, KRAS in colorectal cancer) [16]. Biomarkers and the companion diagnostic that measures them guide decision making, attempting to avoid administering an ill-suited therapy. The FDA is responsible for approving all biomarkers through a rigorous qualification process and maintains a list of approved devices to enable the decision, highlighting the corresponding biomarker and associated disease (aka indication) where the biomarker and the companion diagnostic device are relevant.
Similarly, insurance companies are developing parametric insurance products that calibrate contracts using climate-based indices [17]. Because payouts are triggered when climate index thresholds are met or exceeded, they occur almost immediately without manual claims assessments, which is critical to rapid recovery. In the same way clinical biomarkers can guide a physician’s decision making, climate indices like wind speed, flood height, and earthquake intensity can dictate graduated payouts, provided premiums and payouts are calibrated accurately. Basis risk occurs when there is a mismatch between the loss a policyholder experiences and the final payout from an insurer, and while it exists in traditional insurance, it plays a larger role in parametric insurance [18].
One example of modeling the linkage between property loss and changing climate parameters was reported using atmospheric rivers as an example [19]. Here, property losses due to atmospheric rivers were predicted based on different emission models, providing a path to an integrated view of natural catastrophes and property loss using different prospective emission models. Figure 5 illustrates how the range of predicted annual financial loss varies substantially based on different emission level model predictions.
Science-to-Data-to-Decisions: New Modalities & Frameworks
Parametric insurance is an example of how scientific modeling can support new business models that mitigate climate change effects. Insurers have been analyzing historical records for decades to create actuarial tables, but as in precision medicine, climate risk modeling is expanding to use new modalities and AI-based frameworks to improve prediction and decision making. These frameworks are demonstrating a new and actionable connection between science and mitigating climate change.
Connecting Science and Economics to Unlock Value
Progress in both medicine and climate risk modeling utilize new measurement technologies and algorithms to extract insights. Combined they can reveal complex interconnections across different scales of observation about diseases and natural catastrophes that are beneficial to individuals and subpopulations united by factors as diverse as geography and genetics. Systems biology has developed complex models of biological systems which can then be used in clinical practice [20].
Multiscale climate risk models have existed for decades, but they are now being augmented with economic variables to assess how well science-based climate mitigations work. In Figure 6, a multiscale framework for agricultural climate change interventions enables a gene-to-farm design of resilient and sustainable crop production systems across regional-to-global scales. The linkage between core plant biology and genetics to plant and field behavior to real-world economic outputs connects science-based climate adaptations with expected economic outcomes.
Mitigating Climate Risk for Farmers Using Remote Sensing
In another example of creating value propositions for parametric insurance, a recent article shows how satellite imagery produces measures of agricultural productivity and crop stress that can inform risk models that minimize economic shock and expensive coping strategies that farmers experience because of adverse weather [21]. Figure 7 compares the difference in consequences experienced by farmers when agricultural risk is managed effectively.
Moreover, satellite images, particularly those that collect multiple spectral bands are capable of collecting important agricultural states and the authors discuss how these predictors could be combined with crop models to enable robust insurance indices [22]. Hyperspectral imagery contains hundreds of adjacent spectral bands that when chosen correctly can distinguish and reveal subpixel constituents, suggesting the ability to augment other satellite imaging modalities like radar and visible imagery with orthogonal information [23], [24].
Towards AI-Driven Decision Making
The development of precision medicine has led to AI-based FDA-approved algorithms entering clinical practice. Recently, a supervised machine learning model, Sepsis ImmunoScore ML, was the first FDA-authorized AI-based software designed to identify patients at risk of sepsis [25]. The random forest model was trained on 2366 patient encounters in the derivation cohort using 22 patient-specific features comprising demographics, vital signs, and laboratory tests measured close to study entry. More recently, using deep learning algorithms, Onc.ai was awarded FDA Breakthrough Device Designation for stratifying patient mortality through serial CT images. Despite being less explainable than other ML models, deep learning models are expected to make significant breakthroughs, particularly using patient imagery.
AI/ML is also advancing across many facets of the continuum between climate physical modeling, risk estimation, and insurance. Historically, insurance has been based on actuarial models that use past losses to allocate risk, but model-based approaches for catastrophe modeling are now invoking scientific and computational frameworks. Figure 8 shows four types of climate risks (e.g., droughts, heat, wildfire, floods) with unique data architectures ingesting both temporal and spatial data sources into AI/ML models. There is, however, continued debate about the utility and explainability of different supervised AI/ML models, with deep learning models delivering better performance than random forest, SVM, and regression-based approaches [26].
Risk models, however, can invoke different assumptions and perspectives. A retrospective review of the literature focusing on climate risk insurance modeling found that only half of the papers incorporated different future emission levels to account for multiple scenarios, i.e., they created a family of risk predictions for different future emission scenarios [14]. The most common forward-looking models were the RCPs (Representative Concentration Pathways) that define multiple trajectories to simulate future climate conditions, depicted in Figure 9 [27].
Precision Climate Health: Parallel Journeys that Meet
Thus far, climate risk for parametric insurance and precision medicine have been discussed as if they don’t intersect. But, in fact, they do, and there is increasing recognition that the adverse health effects of climate change are not uniformly distributed and that, beyond obvious distinctions, climate mitigation techniques should be adapted to different subpopulations.
Individual human resilience to heat stress induced by local temperature increases can depend on factors as diverse as genetics, physiology, and the built environment. Socio-economic factors further disperse the range of heat-related effects suggesting governments adopt precision public health approaches to meet the needs of disparate populations [28], [29]. To extrapolate the losses in labor productivity due to heat stress globally in outdoor industries like agriculture, construction, forestry, and fishing, the authors in [30] concatenate a metric for heat (invoking both temperature and humidity) with an exposure response function that relates heat exposure to labor productivity losses. In summary, they state:
“The relationship among global temperatures and global total labor losses and economic productivity losses is inherently non linear as the background climate state changes and the geographic extent of heat exposure increases.”
Mitigation efforts to address rising heat have largely focused on shifting work hours and introducing work breaks, but the former can also lead to less sleep and may introduce noise in early hours that conflict with local ordinances [30]. As such, mitigation efforts need to consider many factors to minimize direct and collateral impacts.
Keys to Unlocking Value for Parametric Insurance
As precision medicine has progressed from discoveries in biology towards clinical applications integrated into insurance frameworks, parametric insurance is gaining traction as climate-driven natural disasters increase and insurance companies seek to recalibrate their business models. To that end, there are opportunities to accelerate this trend.
Minimizing Basis Risk Using Composite Indices
Biomarkers and climate indices are both observables with clear correspondence to biology and science, but, alone, their predictive power may not be sufficiently accurate to create compelling value propositions. Combining multiple variables, however, may improve performance, possibly at the expense of explainability and traceability, and composite biomarkers have been on the radar of the digital health advocates and the FDA [31]. Using multiple, or composite, indices for parametric insurance is frequently viewed as a logical option to minimize basis risk [32]. Whether the composite indices are generated individually and then integrated or embedded directly into a deep learning model, they both represent the same foundational concept: multiple types of data being systematically combined into one multivariate architecture.
Particularly in medicine, however, there is resistance to composite indices that are not easily interpretable and don’t corroborate and reinforce the intuition of doctors and patients. In climate risk modeling, creating a better parametric insurance model through composite indices would likely meet less resistance, but cultural resistance to abstract mathematical models will always remain a challenge in most industries. Tools and methods to improve the explainability of AI/ML models can be especially useful here.
More Remote Sensing Capacity & Resolution Creates Inflection Points
Since the launch of the first commercial remote sensing satellite, Ikonos, in 1999, the number of satellites launched has risen dramatically. As an example of this progression, advanced modalities like hyperspectral imaging (HSI) were only being prototyped and tested by government organizations in 2000, but there are now multiple HSI commercial constellations being launched to augment existing optical, radar, and IR systems. As progress in medical imaging has discovered novel biomarkers and created altogether new treatment pathways, greater temporal and spatial resolution can unlock new business opportunities that were unreachable before.
Increased Accessibility and Customizability
Due to advances in miniaturization and launch cost, remote sensing platforms have become within reach to a far broader class of stakeholders, including smaller countries, companies, and municipalities. And, with the advent of 10 cm-cubed CubeSat frameworks, new platforms can be designed to match user needs with low cost and high versatility. This trend will continue and will accelerate innovation among small companies seeking to address challenges in climate risk modeling and parametric insurance. Near real-time processing, increased downlink speeds, and integration with terrestrial data sources will also be major factors for addressing new gaps and improving performance for existing challenges.
Transparency
Finally, in contrast to the highly regulated environment of medical devices and drug R&D, there has been a lack of transparency among insurance companies about their proprietary catastrophe models. With the exception of one company, First Street, which publishes detailed methodologies, there is a dearth of public-facing information about catastrophe models, making state regulatory reviews one of the few opportunities for the public to scrutinize and insurance models and their performance [18]. The insurance industry, and the public-at-large, could all benefit from greater transparency and forthrightness about how their climate models generate premiums for their customers.
Summary
In this brief survey, we examined the parallel development of climate risk modeling for parametric insurance and precision medicine and highlighted their common foundations and differences. The main observations are:
Parametric insurance can become a massive lever for mitigating and adapting to climate-related risks. Just as biomarkers have created a robust foundation for precision medicine that is reshaping clinical care, climate risk indices, combined with AI/ML, can reallocate risk that is may otherwise be “one-size-fits-all.”
Fidelity between science-driven natural catastrophe frameworks and risk/insurance models is a “moving target” because of climate change but benefits from considering multiple emission scenarios. Appending orthogonal modalities, such as remote sensing imagery, to existing insurance data models can reduce the latency of risk updates that increase incentives for climate adaptation.
Unique AI/ML frameworks for natural catastrophe modeling are important steps but increasingly, models must be able to model and predict multiple hazards such as when Hurricane Helene was followed by flooding in Georgia and North Carolina.
Greater use case diversity and more geographies will accelerate climate adaptation more broadly. Wildfires and flood in North America and Europe have dominated most of the research, leaving other problems and territories largely unstudied.
Increasing resolution, and capacity are opening new value propositions enabled by remote sensing. New constellations with increasing spatial resolution will result in higher revisit rates and more persistence. Inevitably, these new capabilities will improve our ability to monitor existing risks and enable us to address previously unobservable climate threats.
More transparency in the insurance industry about their climate risk models can increase how well consumers and companies can choose solutions that fit their needs.
References
[1] “World Bank Group.” [Online]. Available: https://data.worldbank.org/
[2] “US Health Statistics and Data Trends: life expectancy, health insurance, and more,” USAFacts. Accessed: Jun. 02, 2025. [Online]. Available: https://usafacts.org/topics/health/
[3] A. B. Smith, “U.S. Billion-dollar Weather and Climate Disasters, 1980 - present (NCEI Accession 0209268).” NOAA National Centers for Environmental Information, 2020. doi: 10.25921/STKW-7W73.
[4] “Climate Change and Social Vulnerability in the United States: A Focus on Six Impacts.,” EPA, EPA 430-R-21-003, Sep. 2021. [Online]. Available: https://www.epa.gov/system/files/documents/2021-09/climate-vulnerability_september-2021_508.pdf
[5] “Climate change is showing its claws: The world is getting hotter, resulting in severe hurricanes, thunderstorms and floods.” Munich Re, Jan. 09, 2025. [Online]. Available: https://www.munichre.com/en/company/media-relations/media-information-and-corporate-news/media-information/2025/natural-disaster-figures-2024.html
[6] “HER2+ Breast & Stomach Cancer Treatment Option | Herceptin® (trastuzumab).” Accessed: May 30, 2025. [Online]. Available: https://www.herceptin.com/
[7] N. J. Schork, “Personalized medicine: Time for one-person trials,” Nature, vol. 520, no. 7549, pp. 609–611, Apr. 2015, doi: 10.1038/520609a.
[8] G. Daniel and M. Romine, “The significance of President Obama’s Precision Medicine Initiative.” [Online]. Available: https://www.brookings.edu/articles/the-significance-of-president-obamas-precision-medicine-initiative/
[9] E. Lundstad et al., “The global historical climate database HCLIM,” Sci. Data, vol. 10, no. 1, p. 44, Jan. 2023, doi: 10.1038/s41597-022-01919-w.
[10] “Global Human Settlement - Atlas of the Human Planet 2024 - European Commission.” Accessed: May 30, 2025. [Online]. Available: https://human-settlement.emergency.copernicus.eu/atlas2024.php
[11] T. L. Apgar and C. R. Sanders, “Compendium of causative genes and their encoded proteins for common monogenic disorders,” Protein Sci., vol. 31, no. 1, pp. 75–91, Jan. 2022, doi: 10.1002/pro.4183.
[12] “An introduction to flood modeling for catastrophe risk management”, doi: 10.1002/wat2.1568.
[13] S. Surminski, “The Role of Insurance in Reducing Direct Risk - The Case of Flood Insurance,” Int. Rev. Environ. Resour. Econ., vol. 7, no. 3–4, pp. 241–278, Dec. 2014, doi: 10.1561/101.00000062.
[14] M. W. Ingels, W. J. W. Botzen, J. C. J. H. Aerts, J. Brusselaers, and M. Tesselaar, “The state of the art and future of climate risk insurance modeling,” Ann. N. Y. Acad. Sci., vol. 1541, no. 1, pp. 100–114, 2024, doi: 10.1111/nyas.15255.
[15] C. Lesk et al., “Compound heat and moisture extreme impacts on global crop yields under climate change,” Nat. Rev. Earth Environ., vol. 3, no. 12, pp. 872–889, Dec. 2022, doi: 10.1038/s43017-022-00368-8.
[16] G. A. Lin, J. M. Coffman, and K. A. Phillips, “The State of State Biomarker Testing Insurance Coverage Laws,” JAMA, vol. 331, no. 22, pp. 1885–1886, Jun. 2024, doi: 10.1001/jama.2024.6058.
[17] “How parametric insurance is building climate resilience | World Economic Forum.” Accessed: May 30, 2025. [Online]. Available: https://www.weforum.org/stories/2025/01/what-is-parametric-insurance-and-how-is-it-building-climate-resilience/
[18] J. Cummings and D. Dusseau, “Rebuilding Insurance for a Climate Future,” May 2025.
[19] T. W. Corringham, J. McCarthy, T. Shulgina, A. Gershunov, D. R. Cayan, and F. M. Ralph, “Climate change contributions to future atmospheric river flood damages in the western United States,” Sci. Rep., vol. 12, no. 1, p. 13747, Aug. 2022, doi: 10.1038/s41598-022-15474-2.
[20] L. V. Schaffer and T. Ideker, “Mapping the multiscale structure of biological systems,” Cell Syst., vol. 12, no. 6, pp. 622–635, Jun. 2021, doi: 10.1016/j.cels.2021.05.012.
[21] E. Benami et al., “Uniting remote sensing, crop modelling and economics for agricultural risk management,” Nat. Rev. Earth Environ., vol. 2, no. 2, pp. 140–159, Jan. 2021, doi: 10.1038/s43017-020-00122-y.
[22] D. B. Lobell, D. Thau, C. Seifert, E. Engle, and B. Little, “A scalable satellite-based crop yield mapper,” Remote Sens. Environ., vol. 164, pp. 324–333, Jul. 2015, doi: 10.1016/j.rse.2015.04.021.
[23] N. Keshava, “Distance metrics and band selection in hyperspectral processing with applications to material identification and spectral libraries,” IEEE Trans. Geosci. Remote Sens., vol. 42, no. 7, pp. 1552–1565, Jul. 2004, doi: 10.1109/TGRS.2004.830549.
[24] N. Keshava and J. F. Mustard, “Spectral unmixing,” IEEE Signal Process. Mag., vol. 19, no. 1, pp. 44–57, Jan. 2002, doi: 10.1109/79.974727.
[25] A. Bhargava et al., “FDA-Authorized AI/ML Tool for Sepsis Prediction: Development and Validation,” NEJM AI, vol. 1, no. 12, p. AIoa2400867, Nov. 2024, doi: 10.1056/AIoa2400867.
[26] G. Camps-Valls et al., “Artificial intelligence for modeling and understanding extreme weather and climate events,” Nat. Commun., vol. 16, no. 1, p. 1919, Feb. 2025, doi: 10.1038/s41467-025-56573-8.
[27] D. P. van Vuuren et al., “The representative concentration pathways: an overview,” Clim. Change, vol. 109, no. 1, p. 5, Aug. 2011, doi: 10.1007/s10584-011-0148-z.
[28] A. S. Bernstein, J. R. Malits, and G. A. Wellenius, “The Untapped Potential of Precision Health to Improve Heat Resilience for Those Who Need It Most,” Am. J. Public Health, vol. 114, no. 7, pp. 682–684, Jul. 2024, doi: 10.2105/AJPH.2024.307688.
[29] A. Bouchama et al., “Classic and exertional heatstroke,” Nat. Rev. Dis. Primer, vol. 8, no. 1, p. 8, Feb. 2022, doi: 10.1038/s41572-021-00334-6.
[30] L. A. Parsons, D. Shindell, M. Tigchelaar, Y. Zhang, and J. T. Spector, “Increased labor losses and decreased adaptation potential in a warmer world,” Nat. Commun., vol. 12, no. 1, p. 7286, Dec. 2021, doi: 10.1038/s41467-021-27328-y.
[31] J. A. Beaver et al., “An FDA Perspective on the Regulatory Implications of Complex Signatures to Predict Response to Targeted Therapies,” Clin. Cancer Res., vol. 23, no. 6, pp. 1368–1372, Mar. 2017, doi: 10.1158/1078-0432.CCR-16-1098.
[32] T. Dalhaus, O. Musshoff, and R. Finger, “Phenology Information Contributes to Reduce Temporal Basis Risk in Agricultural Weather Index Insurance,” Sci. Rep., vol. 8, no. 1, p. 46, Jan. 2018, doi: 10.1038/s41598-017-18656-5.
 | A guest post by
|