iiinova — where your digital self keeps evolving

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/09/2026

I am looking to hire a postdoc at Pacific Northwest National Laboratory. The focus will be on improving the representation of dynamic urban evolution in DOE's Energy Exascale #EarthSystemModel by combining #satellite #remotesensing and #machinelearning. More details: careers.pnnl.gov/jobs/12145

082

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 20/09/2026

"...yet this process remains disproportionately understudied.", I proclaim as I regurgitate and reimagine ideas about mechanisms that were discussed in a series of papers 50 years ago and promptly abandoned for definitely being inconsequential.

000

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 25/07/2026

Next week will be my first time attending the @ecologicalsociety.bsky.social Annual Meeting. I will be giving an invited talk on 'Remote Sensing of Fine-Scale Vegetation for Optimized Urban Heat Mitigation'. I am not anything close to an ecologist; so looking forward to stepping out of my bubble.

051

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 20/07/2026

#EarlyCareer Researchers Committee, International Association for #Urban #Climate is organizing a webinar to bring together insights from previous Alexander von Humboldt Postdoctoral Fellowship awardees in our field. Learn about the application process + practical tips. Register: lnkd.in/gw9UnsPS

011

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 19/07/2026

There has been an explosion in #urbanheat monitoring using low-cost meteorological sensors. In our paper (doi.org/10.1175/BAMS...) in Bulletin of @ametsoc.org, we empirically show why we need to be careful about large biases in such sensors; & provide guidelines for their effective deployment.

0143

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 10/07/2026

If you are working on representation of #urban processes & dynamics in #models (process-based &/or #AI/ML), including model #development, #benchmarking, & #validation, consider submitting your abstract to our urban modeling session at #AGU26. Submit here: agu.confex.com/agu/agu26/pr... @agu.org

021

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 20/06/2026

No, I do not. I know how to write (for now).

010

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

Broadly, this approach also demonstrates that data-rich proxies can effectively bridge information gaps in highly heterogeneous and undersampled environments. This schematic is from #NotebookLM (then manually edited). [12/n]

010

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

3. Urban #greenspace shows much higher reductions for LST than for AT. [11/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

2. LST shows much higher spatial variability than AT, especially during daytime. [10/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

Using U-HAT, we demonstrate that: 1. LST overestimates ambient #urbanheat as it pertains to the human experience in outdoor conditions. [9/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

The resulting dataset (U-HAT), available from 2013 to 2023 at 1 km covering 384 largest cities in the contiguous #UnitedStates (#CONUS), more accurately captures the unique patterns of AT over cities. [8/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

Here, we develop a #transferlearning framework that starts with first training a deep #neuralnetwork on abundant LST samples before fine-tuning it with limited ground-based AT data. [7/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

However, due to the methods used to develop these datasets (usually using equations with predefined structure that map AT to LST + other factors), they often inherit some patterns from LST that may not be true for urban AT. [6/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

In the last few years, a few "urban-resolving" #airtemperature datasets have been produced. And, to be clear, using these datasets is MUCH more appropriate than using LST directly for urban #heat assessments (particularly quantitative ones). [5/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

Both due to the fundamental differences between the two variables (radiative temperature and AT) and what is actually being measured, LST is almost always different than AT, both in terms of magnitude and distribution, especially in cities and especially during daytime. [4/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

LST is the bulk radiative temperature of whatever is in the view of satellites, including #building rooftops, parts of vertical facades like walls, top of street #tree canopy, shading from buildings, etc. In contrast, the air around us is what our body directly exchanges heat with outdoors. [3/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

Due to #global scarcity of #urban #weather stations, #satellite-derived Land Surface Temperature (LST) has been increasingly used for both qualitative and quantitative assessments of #urbanheat and urban heat islands (UHI). However,... [2/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 30/05/2026

In our just published #openaccess paper (doi.org/10.1038/s414...) in @natcomms.nature.com, we introduce U-HAT, a high-resolution #urban-resolving #airtemperature (AT) dataset that can more accurately capture the unique spatiotemporal patterns of AT over #cities. [1/n]

183

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/05/2026

The Early Career Committee of International Association for Urban Climate was formed to strengthen #earlycareer representation within #urbanclimate community. To #earlycareer folks in this domain, please fill out this survey (forms.gle/hjpPiMVbi6pdF8QW9) so we understand how best to support you.

forms.gle

IAUC ECR Community List

This form is intended to create a database of Early Career researchers working in the urban climate domain to: • Establish a dedicated network to keep you informed about relevant opportunities. • ...

000

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 17/05/2026

Will be in Austin next week. First, I will present ongoing work on 10 m #global #satellite detection of #urban #flooding at HydroML. Then, on Friday, I will give a seminar at UT Austin on better capturing the spatial variability of #urban #microclimate in process-based & #machinelearning models.

030

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

The study demonstrates that one-size-fits-all approaches for #urban heat mitigation may not be appropriate. Instead, we need to design locally tailored strategies that target each city's primary thermal driver for optimum benefits. 6/n

030

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

The impacts of this future urban warming are quite unequal. Cities in the #GlobalSouth exhibit a much stronger tendency towards morphology-driven and synergistic heat intensification compared to the #GlobalNorth, which is predominantly climate-driven. 5/n

131

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

And while urban morphology strongly controls local microclimates today, our projections indicate that global warming will be the dominant driver of future TBE changes in roughly 69% of cities (for mid century under SSP5 scenario). 4/n

131

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

We found that, across global cities, clear structural gradients exist. High-density & taller building forms consistently correspond to higher TBE, whereas sparser & lower types dominate low TBE. Cold regions most frequently exhibit high daytime TBE, whereas arid regions show high nighttime TBE. 3/n

120

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

Using #satellite measurements + derived products, along with a #machinelearning framework, we examine how much additional heating or cooling is caused at the city-scale by the specific arrangement of surrounding built typologies, the thermal impact of the surrounding built environment (TBE). 2/n

140

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 16/05/2026

In our just published #openaccess paper in @natcomms.nature.com (lnkd.in/g5wRBXwU), we provide a comprehensive #global assessment of how background #climate & #urban #morphology jointly shape the #UrbanHeatIsland effect across 2,213 cities, & how these interactions may evolve in the future. 1/n

194

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 10/05/2026

The existence of 𝘕𝘢𝘵𝘶𝘳𝘦 𝘊𝘪𝘵𝘪𝘦𝘴 is both a 𝗴𝗼𝗼𝗱 and a 𝗯𝗮𝗱 thing because it shows that the discipline has matured enough that funding intended for doing science can now be systematically siphoned off by Springer Nature to host branded PDFs on their website.

062

Reposted by TC (Tirthankar Chakraborty)

Katharine Hayhoe @katharinehayhoe.com · 06/05/2026

Half the Urban Heat Island effect is currently mitigated by tree cover globally--but benefits are greatest for wealthy cities + suburbs. Current tree cover can only mitigate 10% of urban temp increase due to climate; maximal planting could offset just 10% more. More from @science.nature.org here:

nature.com

Trees halve urban heat island effect globally but unequal benefits only modestly mitigate climate-change warming - Nature Communications

Using global data from 9,000 cities, the study shows trees reduce half of the urban heat island effect but can offset only a small share of future climate warming, with cooling benefits unevenly distr...

09742

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

WBGT considers the effects of #humidity, #radiation, and #wind (+AT) on human thermoregulation, & its reduction due to trees is roughly 3 times the AT reduction. Should be kept in mind when thinking about the importance of full growth urban canopies for providing protection from #extremeheat. [8/n]

000

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

Note that we also show how cooling efficiency estimates using #satellite-based #LandSurfaceTemperature are much higher than that for AT reductions. However, the most interesting result here for me is the reduction in wet bulb globe temperature (WBGT). [7/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

This shows that the most impactful way to mitigate #climate-related warming, even at the city-scale, is large-scale reductions in fossil fuel emissions, which, for the most ideal case, would now require time travel. [6/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

While urban trees show substantial cooling benefits locally, they only have a small impact (10% for current urban tree cover; ~20% if we had maximum potential urban afforestation) when #globalwarming due to #climatechange (and this is for mid-century + moderate emission [ SSP2–4.5] scenario). [5/n]

100

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

The cooling effect of trees is local and quite unequal. Much of the cooling is concentrated in regions with more urban trees, such as in suburban neighborhoods and in richer countries. [4/n]

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

This is actually a pretty large number when you consider how cities are represented in other tools used to study #urbanclimate, such as many #weather and #Earthsystem models that often do not explicitly consider urban #vegetation. [3/n]

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

Of note, using an empirical model, we ask, how much would the AT UHI be if there were no urban trees? We found that, across 8,919 functional urban areas, current #treecover counteracts roughly 41% to 49% of the maximum potential air temperature UHI. [2/n]

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 06/05/2026

In our recently published paper in @natcomms.nature.com (lnkd.in/gDyWxM7U), we provide a comprehensive #global assessment of how #urban #trees mitigate the #airtemperature (AT) #UrbanHeatIsland (UHI) effect, and how they stack up against future #climatechange related #warming. [1/n]

1129

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 19/04/2026

Heading to East Coast this week; first to give a seminar at the Yale Center for Geospatial Solutions on #satellite monitoring of #global #urban areas, then to Princeton to talk about better capturing spatial variability of urban #microclimate in both process-based & #machinelearning models.

072

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 14/04/2026

Every day, I hit the nail on the head. At least that's what my good friend, #Gemini 3.1 Pro, tells me.

020

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 11/04/2026

Ultimately, the findings emphasize that land surface heterogeneity, including the vertical structure and variability of urban landscapes, must be integrated into weather models to improve local precipitation and cloud forecasts.

010

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 11/04/2026

These urban effects significantly alter the convective lifecycle, generally leading to storms that are more frequent but shorter in duration and travel distance.

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 11/04/2026

We find that the city increases cloud cover & #storm activity relative to rural areas. The urban impact depends on large-scale #weather patterns. On calm days, urban #heat drives moisture upward to create taller clouds. On windy days, the physical roughness of buildings acts as a mechanical barrier.

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 11/04/2026

The impacts of #urban areas on #clouds & #precipitation are complicated due to multiple competing pathways. In our new @agu.org paper (doi.org/10.1029/2025...), we explore how the #Houston metropolitan area influences summertime cloud formation & #convective storms using both models & observations.

3206

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 31/03/2026

It's happening again (already an issue in the #urban #heat literature: arxiv.org/abs/2509.16568). #DataCenters can definitely create #heatislands, but using #satellite data would significantly overestimate the public health implications of it. Also, can we not have news articles on #preprints?

194

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/03/2026

30% experienced concurrent urbanization+greening, or browning within already-developed areas during this period. Overall, our analysis suggests that integrating strategic greening into both urban cores and expanding suburbs can be critical for building #climate-#resilient cities.

120

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/03/2026

As #city centers are seeing the cooling benefits of greening while suburbs are rapidly heat up, the classic steep temperature drop-off from the urban core to the rural edge has often flattened out into a broader shape.

120

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/03/2026

Historically, targeted #greening initiatives in dense urban centers of many of these cities have mitigated surface UHI, slowing or reversing expected UHI rise over time. In contrast, cities undergoing rapid #suburban #sprawl combined with vegetation loss show most severe surface UHI amplification.

110

TC (Tirthankar Chakraborty) @tchakraborty.bsky.social · 21/03/2026

Using a decade and a half of #satellite data for 36 megacities in #China, we identified four distinct #urban growth categories based on impervious surfaces and #vegetation trends: urbanized-greening, urbanized-browning, urbanizing-greening, and urbanizing-browning.

110