class: center, middle, inverse, title-slide # Downwind and Out: The Strategic Dispersion of Power Plants and Their Pollution ## SBCA, March 2021 ### John M. Morehouse & Edward Rubin --- <style type="text/css"> @media print { .has-continuation { display: block; } } </style> # Air Quality Regulation Air-quality regulation in the US has typically followed a federalist approach. The Clean Air Act of 1963 (and subsequent amendments): --- count: false # Air Quality Regulation Air-quality regulation in the US has typically followed a federalist approach. The Clean Air Act of 1963 (and subsequent amendments): - __Federal__ agencies set national ambient air quality standards (NAAQS) - __State__ governments enforce NAAQS (setting implementation plans, among other things) - __Local__ governments monitor air quality and participate in siting polluters/monitors --- count: false # Air Quality Regulation Air-quality regulation in the US has typically followed a federalist approach. The Clean Air Act of 1963 (and subsequent amendments): - __Federal__ agencies set national ambient air quality standards (NAAQS) - __State__ governments enforce NAAQS (setting implementation plans, among other things) - __Local__ governments monitor air quality and participate in siting polluters/monitors .red_b[Problem]: Air pollution can travel long distances and not all counties are monitored - Regulation & enforcement are complicated! --- # Our paper This paper has 3 main goals: --- count: false # Our paper This paper has 3 main goals: **1.** Describe the geography of a major class of polluters: power plants --- count: false # Our paper This paper has 3 main goals: **1.** Describe the geography of a major class of polluters: power plants **2.** Identify reasons (both strategic _and_ non-strategic) for observed patterns --- count: false # Our paper This paper has 3 main goals: **1.** Describe the geography of a major class of polluters: power plants **2.** Identify reasons (both strategic _and_ non-strategic) for observed patterns **3.** Illustrate the extent of the pollution-transport problem --- count: false # Our paper This paper has 3 main goals: **1.** Describe the geography of a major class of polluters: power plants **2.** Identify reasons (both strategic _and_ non-strategic) for observed patterns **3.** Illustrate the extent of the pollution-transport problem .red_b[Why?] Air-pollution regulation and monitoring is fraught with complexity. We shed light on additional challenges regulators face under the current, federalist system. --- # Literature In general, our work is related to two strands of literature: __Strategy and the CAA__ --- count: false # Literature In general, our work is related to two strands of literature: __Strategy and the CAA__ + Downwind siting for polluters as a strategy (_e.g._ Monogan III et. al (2017)) + Strategic abatement decisions (_e.g._ Zou, 2020) + Strategic _monitor_ placement (_e.g._ Grainger et. al, 2018) --- count: false # Literature In general, our work is related to two strands of literature: __Strategy and the CAA__ + Downwind siting for polluters as a strategy (_e.g._ Monogan III et. al (2017)) + Strategic abatement decisions (_e.g._ Zou, 2020) + Strategic _monitor_ placement (_e.g._ Grainger et. al, 2018) __Pervasiveness and problems with pollution transfer__ + Sergi et. al (2020), Wang et. al (2020), Tessum et. al (2017) + Quantify extent of pollution transport in general + costs (health damages) --- class: middle, inverse, center # The Geography of US Power Plants --- # Data Sources .red_b.purple[Generator Data:] `Emissions & Generation Integrated Database (eGRID)` and `EPAs EmPOWER Air Data Challenge`. --- count: false # Data Sources .red_b.purple[Generator Data:] `Emissions & Generation Integrated Database (eGRID)` and `EPAs EmPOWER Air Data Challenge`. .red_b.purple[Geography:] - `US Census Bureau Tiger/Line` shapefiles for county, state, and water features. - `EPA's Greenbook NAYRO` file for county non-attainment histories --- count: false # Data Sources .red_b.purple[Generator Data:] `Emissions & Generation Integrated Database (eGRID)` and `EPAs EmPOWER Air Data Challenge`. .red_b.purple[Geography:] - `US Census Bureau Tiger/Line` shapefiles for county, state, and water features. - `EPA's Greenbook NAYRO` file for county non-attainment histories .red_b.purple[Meteorology:] `NOAA's North American Regional Reanalysis` (NARR) daily data Historic wind patterns at various pressure levels. 32km `\(\times\)` 32km grid cells across contigous US --- # Distances to County Borders <img src="final_figs/county_dist_1.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- count: false # Distances to County Borders <img src="final_figs/county_dist_2.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- # Distances to State Borders <img src="final_figs/state_dist_1.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- count: false # Distances to State Borders <img src="final_figs/state_dist_2.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- # Water Borders: Example .col-left[ <img src="figures/borders-water-41.png" width="80%" height="80%" style="display: block; margin: auto;" /> <img src="figures/borders-water-22.png" width="80%" height="80%" style="display: block; margin: auto;" /> ] .col-right[ <img src="figures/borders-water-46.png" width="80%" height="80%" style="display: block; margin: auto;" /> <img src="figures/borders-water-45.png" width="80%" height="80%" style="display: block; margin: auto;" /> ] --- # A test for regulatory avoidance We can't say strategy .red_b[caused] border siting. --- count: false # A test for regulatory avoidance We can't say strategy .red_b[caused] border siting. .red_b[Question:] Do power-plants (excluding maybe wind) use the ratio of upwind/downwind area within their own county/state to produce electricity? --- count: false # A test for regulatory avoidance We can't say strategy .red_b[caused] border siting. .red_b[Question:] Do power-plants (excluding maybe wind) use the ratio of upwind/downwind area within their own county/state to produce electricity? - Seems unlikely - This is the basis for our identification strategy. --- count: false # A test for regulatory avoidance We can't say strategy .red_b[caused] border siting. .red_b[Question:] Do power-plants (excluding maybe wind) use the ratio of upwind/downwind area within their own county/state to produce electricity? - Seems unlikely - This is the basis for our identification strategy. - Why would a smaller downwind area within a county be advantageous for a polluter? .red_b[Emissions will exit the jurisdiction faster]. --- count: false # A test for regulatory avoidance We can't say strategy .red_b[caused] border siting. .red_b[Question:] Do power-plants (excluding maybe wind) use the ratio of upwind/downwind area within their own county/state to produce electricity? - Seems unlikely - This is the basis for our identification strategy. - Why would a smaller downwind area within a county be advantageous for a polluter? .red_b[Emissions will exit the jurisdiction faster]. __Main Idea:__ In the absence of regulatory avoidance, it should be a 50-50 flip whether the county’s area downwind of the plant (in the EGU’s county of residence) is larger or smaller than the area upwind. - .red_b[Focus]: coal. Strongest incentive to avoid regulation. - .red_b[Placebo]: natural gas. Less incentive to avoid regulation. --- # Downwind vs. Upwind Area <img src="final_figs/upwind_downwind_1.png" width="100%" height="100%" style="display: block; margin: auto;" /> --- # Formalizing the test Our test is implemented via a .red_b[Fisher's exact test] .col-left[ - Sharp Null: _no strategic siting to reduce downwind area_ ] --- count: false # Formalizing the test Our test is implemented via a .red_b[Fisher's exact test] .col-left[ - Sharp Null: _no strategic siting to reduce downwind area_ - Test stat `\(n_s\stackrel{H_0}{\sim} B(N_T,.5)\)` - `\(n_s\)`: # plants for whom downind area `\(<\)` upwind area - `\(N_T\)`: total # plants (within fuel type) - `\(p(n_s) =\sum\limits_{x = n_s}^{N^T}{N_T \choose x}\times 0.5^{N_T}\)` ] --- count: false # Formalizing the test Our test is implemented via a .red_b[Fisher's exact test] .col-left[ - Sharp Null: _no strategic siting to reduce downwind area_ - Test stat `\(n_s\stackrel{H_0}{\sim} B(N_T,.5)\)` - `\(n_s\)`: # plants for whom downind area `\(<\)` upwind area - `\(N_T\)`: total # plants (within fuel type) - `\(p(n_s) =\sum\limits_{x = n_s}^{N^T}{N_T \choose x}\times 0.5^{N_T}\)` ] .col-right[ ➕ Simple and plausible identifiying assumption ➕ Calculate _exact_ p-values. No parameteric assumptions required! ➕ Convenient falsificaton test: Natural gas ] --- count: false # Formalizing the test Our test is implemented via a .red_b[Fisher's exact test] .col-left[ - Sharp Null: _no strategic siting to reduce downwind area_ - Test stat `\(n_s\stackrel{H_0}{\sim} B(N_T,.5)\)` - `\(n_s\)`: # plants for whom downind area `\(<\)` upwind area - `\(N_T\)`: total # plants (within fuel type) - `\(p(n_s) =\sum\limits_{x = n_s}^{N^T}{N_T \choose x}\times 0.5^{N_T}\)` ] .col-right[ ➕ Simple and plausible identifiying assumption ➕ Calculate _exact_ p-values. No parameteric assumptions required! ➕ Convenient falsificaton test: Natural gas ➖ Major drawback: cannot capture more nuanced strategy ] --- # Strategic Siting: Main Results <img src="final_figs/table_header.png" width="80%" height="80%" style="display: block; margin: auto;" /><img src="final_figs/panel_a.png" width="80%" height="80%" style="display: block; margin: auto;" /> --- count: false # Strategic Siting: Main Results <img src="final_figs/table_header.png" width="80%" height="80%" style="display: block; margin: auto;" /><img src="final_figs/panel_b.png" width="80%" height="80%" style="display: block; margin: auto;" /> --- class: middle, inverse, center # The Geography of US Coal Emissions --- # Overview We quantify the nature of the pollution transfer problem --- count: false # Overview We quantify the nature of the pollution transfer problem. __Model__: __HY__brid __S__ingle-__P__article __L__agrangian __I__ntegrated __T__rajectory (HYSPLIT) - Atmospheric dispersion model. Heavily vetted by NOAA. - Performs better than many other models (such as InMAP) for _long-distance_ pollution transport modeling. --- count: false # Overview We quantify the nature of the pollution transfer problem. __Model__: __HY__brid __S__ingle-__P__article __L__agrangian __I__ntegrated __T__rajectory (HYSPLIT) - Atmospheric dispersion model. Heavily vetted by NOAA. - Performs better than many other models (such as InMAP) for _long-distance_ pollution transport modeling. - Coal-based particles will travel much further than other sources of PM. --- # Hysplit: Goals We do the following: 1) Quantify how quickly coal-based particles leave their own county and state (it's fast). 2) Quantify the proportion of coal-based emissions that are from other counties/states in any given county/state. 3) Illustrate the implications of 1) and 2) with case studies. --- # Example Plants <img src="final_figs/example_plants_1.png" width="90%" height="90%" style="display: block; margin: auto;" /><img src="final_figs/col_bar.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- count: false # Example Plants <img src="final_figs/example_plants_2.png" width="90%" height="90%" style="display: block; margin: auto;" /><img src="final_figs/col_bar.png" width="90%" height="90%" style="display: block; margin: auto;" /> --- # Emissions Transport: Speed <img src="final_figs/speed_1.png" width="100%" height="100%" style="display: block; margin: auto;" /><img src="final_figs/speed_legend.png" width="100%" height="100%" style="display: block; margin: auto;" /> --- # Emissions Transport: Speed <img src="final_figs/speed_2.png" width="100%" height="100%" style="display: block; margin: auto;" /><img src="final_figs/speed_legend.png" width="100%" height="100%" style="display: block; margin: auto;" /> --- # Emissions Transport: Shares <img src="final_figs/hysplit_1.png" width="100%" height="100%" style="display: block; margin: auto;" /><img src="final_figs/hysplit_legend.png" width="100%" height="100%" style="display: block; margin: auto;" /> --- class: middle, inverse, center # Discussion --- # What did we do? __Main contributions__: - Descriptive results on the geography of physical power plants _and_ their emissions. - Causal evidence of coal plants strategically locating to minimize downwind area. - Clean Air Act did not seem to impact strategic siting. - Descriptive results on pervasiveness of pollution transport problem from coal powered plants. --- class: middle, inverse, center # Thank you! email: jmorehou@uoregon.edu web: www.johnmmorehouse.com --- exclude: true