Libby H. Koolik

@libbyhkoolik.bsky.social

134 Followers 188 Following

(she/her) UC Berkeley environmental engineering PhD candidate | MIT '17 and '18 | air quality & equity scientist | amateur vegan chef | https://lkoolik.github.io/

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.
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.
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.
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.
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.
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.
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.
News article screenshot from the New York Times: House Votes to Repeal California’s Clean Truck Policies Lawmakers voted to stop the state from requiring that an increasing share of new trucks sold there have zero emissions.
Two panels demonstrating changes in exposure (left) and relative disparity in exposure (right) across time for demographic groups in California. In the left hand panel, we show that population-weighted average exposures for all demographic groups (including Asian, Black, Hispanic, white, and other, plus two regulatory categories for communities) have decreased across time. In the right hand panel, we show that the relative inequality is essentially constant across time for all groups.
Libby H. Koolik and Chirag Manchanda presenting an academic poster at the Health Effects Institute conference.