July 2011 Record Heat and Its Impact on Air Quality

THE STORY
Source: Maryland Department of the Environment's Quality of Air Summaries. (It is copied and pasted directly here.)

Record heat defines the country in July 2011! Thousands of surface meteorological stations across the Country observed daily maximum temperatures that either tied or broke the old records (top panel maps [on dashboard]). The lower Great Plains (TX, OK, AR, MS) stood out the most with stations observing record temperatures on 10-to-21 out of 31 days, as the region also experienced exceptional drought conditions. At the same time, another cluster of record breaking temperatures occurred across the Great Lakes, Mid-Atlantic and Northeast on as many as 5 days. In addition, scattered areas across the Appalachians had to endure record breaking high temperatures on as many as 10 days. These areas were likely influenced by a persistent lee-side-trough that positioned itself across the I-95 corridor as well as an anomalous high pressure system off the Mid-Atlantic Coast that transported warmer tropical air into these areas. Interestingly, during this record breaking heat, the average temperature difference (as much as 4 to 9 degrees F) between the new and old records was the greatest in the Mid-Atlantic and Northeast and parts of the Great Lakes. In theory, record setting temperatures should have posed a very big challenge for air quality in those areas. So was air quality in July 2011 adversely affected by the extremely high temperatures? To answer this question, 2001-2011 air quality data across the U.S. were extracted from the EPA AIRNow real-time database and plotted (bottom map). In particular, the number of days exceeding 100 AQI (Unhealthy for Sensitive Groups or above) in July 2011 was compared to the 10-year average (2001-2010) in order to get a sense of “normal.” This data comparison showed mixed results. Most Core Based Statistical Areas (CBSA) across the country observed less bad air days in July 2011 as compared to 10-year average. However, there were many areas over the lower Great Plains, Great Lakes, Mid-Atlantic and Northeast observing more bad air days in July 2011 as compared to 10-year average. These areas include the Baltimore-Towson, MD and Washington-Arlington-Alexandria, DC-VA-MD- WV CBSAs. In contrast, the Hagerstown-Martinsburg, MD-WV (part of the Western Maryland Forecast Region) and Seaford, DE (surrogate for Eastern Shore Forecast Region) CBSAs were below the 10-year average. Going back to the regional view, some areas (both rural and urban) observed an increase of several bad air days in July. For many of these areas, the increase was only a few days but still significant compared to previous years. In the midst of all this, it’s important to recognize that effective regional and local pollution control programs implemented during 2002-2004 have resulted in improving air quality. However, record heat in July 2011 proved to be an overwhelming factor which caused poor air quality. Increasing temperature trends have and probably will continue to erode away air quality progress of recent years. The degree of potency remains to be seen but climate forecast and its effects on air quality doesn’t look too promising. See U.S. Global Change Research Program (USGCRP) scientific assessment on regional climate change impacts on the Northeast for details.

INTERACTIVE VISUALIZATION

Use the interactive dashboard below to explore the data and come up with a story in your area. Reload the page if the page appears to be blank. You should see something like this image.

Powered by Tableau

DISCUSSION OF VISUALS

On August 4, the National Oceanic and Atmospheric Administration (NOAA)’s Environmental Visualization Laboratory (EVS) published a visualization to demonstrate the relentless heat which plagued the contiguous U.S. during July 2011. I think the EVS’s map works exceptionally well when the intention is to qualitatively illustrate a sense of magnitude. They achieved this by stripping out all non-data elements (e.g. map background, borders etc.) and only presented the data geographically using a simple choropleth map on a black background. The counts of daily broken records are presented using a sequential color ramp (also a good practice to present geographic data). Refer to Stephen Few’s Visual Business Intelligence Newsletter titled “Introduction to Geographical Data Visualization” for further reading.

The EVS's visualization inspired me to explore how record setting temperatures have impacted the air quality in the Mid-Atlantic and other parts of the country. The focus is daily maximum temperatures as they have a stronger influence over the formation of ground-level ozone, which is one of the pollutants of concern in the U.S. during the summer time.

I started by recreating a map showing the number of broken high temperature records. This task was straightforward using Tableau software with the raw data extracted from the National Climatic Data Center (NCDC) U.S. Records web page. I used the choropleth map approach with slight differences (highlighted below) compared to the EVS’s version. Both versions (in my opinion) work well based on their intentions. My focus was to show the data a bit more quantitatively.

• Used variation of color intensity (shades of red) and size to denote differences in values; Using both color and size helped highlight areas with the greatest impact (e.g. the lower Great Plains and to some extent parts of the Great Lakes, Mid-Atlantic and Northeast). Please note that size legend has been purposely omitted since the color legend should be sufficient for the map.
• Used less color variations to quickly highlight large differences between varying intervals showing the number of broken high temperature records. The eyes are not good at differentiating small differences in color intensity.
• Used unfilled circles to avoid overlapping of data.

The severity of the heat can also be explored in other ways. I potted a map showing the count of triple-digit temperature records and a map showing the average temperature difference between the old vs new record. The triple-digit temperature map showed a similar spatial pattern as compared to the number of broken temperature record map. It would work well if the intention was to highlight the area(s) that endured drought conditions (i.e. lower Great Plains or TX, OK, AR, MS). The averaged temperature difference map on the other hand showed additional insight for the severity of the impact. The visual design elements of this map were the same as those noted above for the number of broken high temperature records map (see above bullets). This map highlighted the Mid-Atlantic, Northeast and parts of the Great Lakes as the areas with the greatest impact due to record setting temperatures. In theory, those areas would have endured the greatest degradation in air quality. This lead to the big question “was air quality in July 2011 adversely affected by the extremely high temperatures?”

To answer this question, 2001-2011 air quality data across the U.S. were obtained and plotted. In particular, the number of bad air days (days exceeding 100 AQI) in July 2011 was compared to the 10-year average (2001-2010) to get a sense of “normal.” This comparison was simple yet able to provide context for understanding the qualitative state of air quality conditions in a particular area. I applied a technique that Mr. Few preaches as one of the good practices for dashboard design. The same technique was applied for the air quality map as well as the time-series graph that show the number of days exceeding 100 AQI. The technique is elaborated below:

• Defined 2001-2011 average as “normal” and computed percent change between 2011 vs 2001-2010 (i.e. departure from normal).
• Defined the qualitative state of 2011. I defined normal as percent change within ± 20% of normal; likewise, values below/above ± 20% would be below/above normal. This is consistent between air quality map and the time-series graph showing the number of days exceeding 100 AQI.
• Shaded ± 20% about the 2001-2010 average to explicitly show normal vs below/above normal conditions in the time-series graph.

Once the graph/maps are designed, publishing them in the form of a “dashboard” was a breeze using Tableau software. Tableau also allows “smart interaction” which Mr. Few often refers to as “brushing and linking.” I used 2 quick filters in my dashboard. The first one linked temperature maps and is used to explore spatial pattern at varying temperature change between old vs new record (e.g. areas of extremes). The second filter linked the air quality map with the time-series graph and is used to quickly filter air quality data by micropolitan and/or metropolitan areas. Unfortunately, due to the nature of the data (both temperature and air quality), those filters can’t be used to filter all maps/graph on the dashboard.

I hope that the discussion of the visuals used in the dashboard to communicate this particular air quality was sufficient.  I hope that it provides enough meaningful comparisons to show how record setting temperatures impacted the air quality in the U.S. during July 2011.  What are your thoughts in regards to the visuals?  Please participate in the discussion to help promote thoughtful presentation of air quality information. This is a learning process.

DISCLAIMER: Any statements, postings, shares, likes, follows on the blog should be considered personal. They do not represent any political or official endorsement from my employer.