Libby H. Koolik

As EPA Revokes Endangerment Finding, Environmental Justice Organizations Continue to Fight to Protect Communities From Climate Chaos & Hold EPA Accountable As EPA Revokes Endangerment Finding, Environmental Justice Organizations Continue to Fight to Protect Communities From Climate Chaos & Hold EPA Accountable

Here is @weact4ej.bsky.social 's response to the current's admirations efforts to ask us to ignore reality so that oil and gas's obscene profits and wealth can grow even larger at the expense of our health, homes, and lives.

Unrelated to the science: this paper has a special place in my heart, as I originally developed this case study with an undergraduate mentee in 2023. With the help of each of our co-authors, it grew into something much more incisive and will hopefully be a useful resource for the field! (9/9)

As the field continues to design policies and interventions that mitigate disparate exposures, we hope that our paper provides a blueprint for the kinds of sensitivity tests and considerations that should go hand-in-hand with future analyses. (8/)

All together, our study highlights why we need to be really intentional when we design equity-oriented air pollution studies. The same dataset can provide a range of answers, so it’s crucial to be clear about what your question is. (7/)

There’s more to unpack than can fit here, but notice the stark differences between the disparities in Agricultural Areas versus the San Joaquin Valley, especially when evaluated relative to statewide versus regional population. It all depends on your policy goals! (6/)

Relative disparity in exposure is calculated for the Hispanic and Black populations over four geographic domains: statewide, within Agricultural Areas, within the San Joaquin Valley, and within Overall High Exposure Areas. Disparities are calculated relative to the statewide average exposure and the regional average exposure. Maps are drawn below with the regions shaded.

We also use four different geographic domains within California to compare how the magnitude of disparity changes when evaluated across different study geographies and reference populations. (5/)

Why does this matter? If we want to develop a policy that targets reducing the disparity in exposure experienced by Black Californians, intervening in the areas most polluted by agricultural sources probably won’t do it. (4/)

Left: Distributions of exposure to PM2.5 from agricultural sources for the total population (black), the Hispanic population (pink), and the Black population (orange). The plot is shaded based on whether it represents exposure lower than the total population (green) or higher (yellow). Right: the “offset” is calculated by comparing the percentile at each exposure concentration of total population and group population distributions. Both: Hispanic people face higher exposures than the total population across the entire distribution. A blue dashed box shows where the Black population changes from higher exposure than the total population to lower exposure.

The first key finding: focusing solely on the population-weighted mean could misrepresent who is burdened with extreme air pollution. We use an “offset” across the distribution to demonstrate the differences between the disparities faced by the Hispanic and Black populations. (3/)

Different terms are defined in bubbles with headers. The key considerations explored in this work are the study geography (geographic extent of study), the exposure input (estimate of exposure used to quantify disparity), and the reference population (group against which disparity is assessed). Within the exposure input, we define exposure distribution (exposure at each functional unit across the domain), the population-weighted mean (average exposure for a group), and distribution point estimates (exposure for a given percentile of a population-weighted distribution). These choices are used to calculate disparity indicators, which include the Atkinson, Gini, Dissimilarity Indices (statistical measures); demographic differences (comparison of relative composition of demographic groups across exposure distribution); exposure gap (difference between most and least exposed groups); absolute disparity (difference between group and total population); and relative disparity (percent difference between group and total population). Finally, these metrics are mapped to potential findings. All but the Indices – which lead to cross-sectional differences in exposure across a population – lead to differences in exposure for a given group that can be linked to unfair and systemic biases. When combined with socially-contextualized and intersectional analysis, these can have the impact of informing solutions that dismantle systemic biases and unfair treatment.

To do this, we first define a few important terms for doing equity-oriented air pollution exposure science. We then explore these terms using a case study of disparities in exposure to modeled air pollution from California’s agricultural sector. (2/)

Methodological Design Choices Can Affect Air Pollution Exposure Disparity Estimates: A Case Study on California’s Agricultural Sector People of color in the United States are disproportionately and unfairly exposed to air pollution. Equity-oriented scientific evaluations quantifying these disparities often use population-average exp...

Out in ES&T (@pubs.acs.org) yesterday: we demonstrate the importance of a few different methodological choices for understanding and ultimately mitigating air pollution exposure disparities. The full article is available at the link below open-access (1/)

pubs.acs.org/doi/10.1021/...

I feel so lucky to have a team of intelligent and kind collaborators on this perspective. Our team represents folks from both academia and EJ community advocates. Thank you all for your support on this work! (10/10)

That’s great for people in states like CA, but not for folks in less EJ-inclined states. My hope is that the good people in local government find our insights helpful and can identify small wins like moving a bus line to reduce traffic in a community or siting EV chargers in polluted areas. (9/)

Still, a key finding of our work is that land use and transportation planners have a really important role in mitigating air pollution exposure disparities. These decisions happen at local and community scales, and are even more important without federal support. (8/)

In the year since I started this project, the regulatory landscape for environmental policy has changed a lot (understatement). Without a federal government pushing for place-based infrastructure changes or emission mitigations, it will be hard to apply this framework at scale. (7/)

PM2.5 exposure disparities persist despite strict vehicle emissions controls in California Decades of California’s strict exhaust emissions controls reduced PM2.5 exposure but not relative racial-ethnic disparity.

The idea was born from this question from my vehicle analysis: why do huge emission reductions not yield equivalent reductions in disparity? What we found was exactly what EJ groups have been advocating for: structural changes to the distribution of emissions sources. (6/)

doi.org/10.1126/scia...

We also apply our conceptual framework to the results that other great scientists in the field have produced to identify key components of successful policies. I am really appreciative of the open science & kindness of these other researchers – please go check out their great publications! (5/)

Our simple model runs three emission-reduction scenarios in space for a simplified city of two linearly-segregated population groups. The first panel introduces the model scenarios. In the first scenario, emissions are reduced in-place. In the second, emissions are reduced while moving away from the city-center but along the same segregation parallel. In the third, emissions are reduced while moving away from both the city center and the overburdened group. The second two panels demonstrate the results of each scenario. In the middle, we show the changes in each variable that occur at the midpoint. The third panel shows the change in absolute disparity as a function of emissions reduction. In both panels, the most substantial reduction in disparity occurs in the third scenario.

In fact, we use a simple Gaussian plume model to demonstrate why we really should design policies that do all three at the same time! Only when we meaningfully relocate the center of mass of emissions do we fully mitigate disparities before eliminating 100% of emissions. (4/)

Four cartoons depicting a highway system running through a segregated community of gold and blue houses. In the “Base Case” in Panel A, the vehicles disproportionately drive through the gold community. The next three panels show the following changes: (B) emissions are reduced but the highway hasn’t moved, (C) the whole highway is shifted uniformly away from all residents, and (D) the off-ramp through the gold community is removed.

Consider this highway system example (A). Some policies reduce emissions in-place (B), some move emissions away from everybody (C), and some meaningfully eliminate the unfair source distribution (D). While real policies do these together, different regulators tend to control each term. (3/)

In our perspective, we introduce a conceptual framework to mitigate exposure disparities that does two things. 1: it decomposes why different policies have different outcomes mathematically. 2: it builds an intuition for which policies should be successful at reducing exposure disparities. (2/)

Eliminating air pollution disparities requires more than emission reduction | PNAS In the United States, people of color are disproportionately and unjustly exposed to air pollution. Historically, environmental policy has emphasiz...

The second chapter of my dissertation was published in @PNAS today! I’m really proud of this perspective, and I want to share a few additional thoughts. The full article is available at the link below open-access (1/)

doi.org/10.1073/pnas...

I feel so lucky to have a team of intelligent and kind collaborators on this perspective. Our team represents folks from both academia and EJ community advocates. Thank you all for your support on this work! (10/10)

That’s great for people in states like CA, but not for folks in less EJ-inclined states. My hope is that the good people in local government find our insights helpful and can identify small wins like moving a bus line to reduce traffic in a community or siting EV chargers in polluted areas. (9/)

Still, a key finding of our work is that land use and transportation planners have a really important role in mitigating air pollution exposure disparities. These decisions happen at local and community scales, and are even more important without federal support. (8/)

In the year since I started this project, the regulatory landscape for environmental policy has changed a lot (understatement). Without a federal government pushing for place-based infrastructure changes or emission mitigations, it will be hard to apply this framework at scale. (7/)

The idea behind this framework was born from my first dissertation chapter: why do massive emission reductions not yield equivalent reductions in disparity? What we found was exactly what EJ groups have been advocating for: structural changes to the distribution of emissions sources. (6/)

We also apply our conceptual framework to the results that other great scientists in the field have produced to identify key components of successful policies. I am really appreciative of the open science & kindness of these other researchers – please go check out their great publications! (5/)

Our simple model runs three emission-reduction scenarios in space for a simplified city of two linearly-segregated population groups. The first panel introduces the model scenarios. In the first scenario, emissions are reduced in-place. In the second, emissions are reduced while moving away from the city-center but along the same segregation parallel. In the third, emissions are reduced while moving away from both the city center and the overburdened group. The second two panels demonstrate the results of each scenario. In the middle, we show the changes in each variable that occur at the midpoint. The third panel shows the change in absolute disparity as a function of emissions reduction. In both panels, the most substantial reduction in disparity occurs in the third scenario.

In fact, we use a simple Gaussian plume model to demonstrate why we really should design policies that do all three at the same time! Only when we meaningfully relocate the center of mass of emissions do we fully mitigate disparities before eliminating 100% of emissions. (4/)

Four cartoons depicting a highway system running through a segregated community of gold and blue houses. In the “Base Case” in Panel A, the vehicles disproportionately drive through the gold community. The next three panels show the following changes: (B) emissions are reduced but the highway hasn’t moved, (C) the whole highway is shifted uniformly away from all residents, and (D) the off-ramp through the gold community is removed.

Consider this highway system example (A). Some policies reduce emissions in-place (B), some move emissions away from everybody (C), and some meaningfully eliminate the unfair source distribution (D). While real policies do these together, different regulators tend to control each term. (3/)

In our perspective, we introduce a conceptual framework to mitigate exposure disparities that does two things. 1: it decomposes why different policies have different outcomes mathematically. 2: it builds an intuition for which policies should be successful at reducing exposure disparities. (2/)

Our key recommendation is still important and possible: we need to prioritize reducing emissions in the most overburdened areas at a rate that outpaces the emissions reductions from the overall on-road fleet turnover. I'm excited to keep looking ahead and thinking creatively about how to do this 👩‍💻