Human Centric Thermal Modeling Beyond Land Surface Temperature

This direction is merged into another direction: surface-to-air temperature discrepancy. Its last version stays readable and citable here.

Version 1 of 126 papers34 conceptsfrozen 2026-10-04

Most climate models and urban planning tools rely on a single, convenient metric to measure heat: Land Surface Temperature (LST). Captured easily by satellites, LST measures the "skin" temperature of the Earth. However, a growing body of research suggests that LST is a poor proxy for the actual thermal experience of living organisms. Whether assessing the cooling benefits of a new city park or the survival of a cranberry crop, researchers are finding that "skin" temperature often fails to capture the complex interplay of humidity, wind, and internal physiology that defines true thermal stress.

The Sensitivity Gap Between Surface and Air

The most fundamental problem with relying on satellite data is the discrepancy between surface temperature ($T_s$) and the air temperature ($T_a$) that humans actually breathe. While a satellite might detect a significant drop in surface temperature after a reforestation project, the cooling felt by a person standing under those trees is often much smaller.

Yitao Li, Zhao-Liang Li, and their colleagues have quantified this "sensitivity gap" across several studies. In The overlooked local air warming mitigation of potential tree restoration, they found that air temperature responses to vegetation changes are typically only about 22% of the surface-level change. Their more recent work, Observed different impacts of potential tree restoration on local surface and air temperature, suggests this figure may be as low as 15% to 30%.

This means that while "greening" the Earth provides a measurable biophysical cooling effect—offsetting nearly 40% of warming in regions like India, as shown in Biophysical impacts of earth greening can substantially mitigate regional land surface temperature warming—the actual relief felt by local populations is significantly lower than what satellite "skin" temperatures suggest.

Why Temperature Alone is Not Enough

If raw temperature is an incomplete metric, what are we missing? The consensus across urban design and public health research is that thermal comfort is multi-modal. Factors like humidity, wind speed, and solar radiation are not just "secondary" variables; they are primary drivers of physiological stress.

In the context of public health, Takuto Sato and Hiroyuki Kusaka demonstrated in Quantitative Assessment of the Contribution of Meteorological Variables to the Prediction of the Number of Heat Stroke Patients for Tokyo that integrating these variables improves heat stroke predictions by 19% compared to temperature-only models. Similarly, Jae-Sik Min and colleagues argue in Assessing Climate-Driven Mortality Risk that standard mortality models underestimate risk because they fail to account for the cumulative impact of humidity and radiation during prolonged heatwaves.

This "soft comfort" or "apparent temperature" approach is also vital for infrastructure. James Doss-Gollin and Yash Amonkar show that using ambient temperature as a sole proxy for energy demand fails to account for wind chill and humidity, which can lead to dangerous underestimations of peak power loads during extreme weather.

Internal States and Biological Reality

The gap between external metrics and internal reality is perhaps most visible in agriculture. In AI on the Bog: Monitoring and Evaluating Cranberry Crop Risk, Benjamin Planche and his team found that the internal temperature of a cranberry can exceed the surrounding air temperature by a staggering 25°C. Relying on air temperature to trigger irrigation systems would result in the fruit burning long before the sensors registered a "danger" zone.

Similarly, Alan Fern and Aseem Saxena used multi-task learning to predict "cold hardiness" in grapevines, discovering that humidity and wind speed were essential for understanding the biological state of the plant—the temperature at which the buds actually freeze—rather than just the temperature of the air around them.

Bridging the Gap with Machine Learning

If the goal is to move from "skin" temperature to "felt" temperature, how do we do it without placing expensive sensors on every street corner or in every bog? The emerging solution is to use deep learning to translate coarse, high-level data into fine-grained, human-centric metrics.

Two notable approaches have recently emerged:

  1. Weakly-Supervised Translation: In Measuring Fine-Grained Urban Air Temperature with Satellite Imagery, Minhyuk Song and Sungwon Han developed the UrbanHeat framework. It uses high-resolution LST as a "proxy label" to train models that eventually predict the more elusive air temperature ($T_a$) at a 0.25km resolution.
  2. Geospatial Foundation Models: Jan H. Fleckenstein and his team utilized foundation models pre-trained on global satellite data to simulate how specific urban interventions—like a new park—would change the microclimate. Their work, Detection and Simulation of Urban Heat Islands, focuses on translating thermal satellite data into metrics that reflect the ground-level experience.

By moving beyond the "skin" of the Earth and integrating the multi-modal variables that define comfort and survival, these researchers are creating a more accurate, human-centric map of our warming world.

Papers in this version

History & citationVersion 1 · 2026-10-04

Merged into another direction · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_human_centric_thermal_modeling_beyond_la_fdc70e49,
  title = {Human Centric Thermal Modeling Beyond Land Surface Temperature},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/human-centric-thermal-modeling-beyond-land-surface-temperature/rdr_fdc70e4964bb41f391602bdec4a705d9},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_human_centric_thermal_modeling_beyond_la_fdc70e49_v1,
  title = {Human Centric Thermal Modeling Beyond Land Surface Temperature},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/human-centric-thermal-modeling-beyond-land-surface-temperature/rdr_fdc70e4964bb41f391602bdec4a705d9/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}