Decrease in (Reported) WMATA Track Delays Since 2009

In a conversation on Twitter, this simple question was tweeted at me:

@MetroReasons @vxgxn @philogrophy @unsuckdcmetro @wmata What's rail incident frequency over time look like?

— Disgruntled Rider (@MetroRage) August 6, 2015

With data from the Daily Service Reports, this shouldn't be too hard to find. Just count up and graph the number of delays categorized under track problems, right? Well, that's exactly what I plan to do with this write-up. If you don't feel like reading everything including the details, I guess you can skip to the bottom and read the conclusion.

Background

The current push for overhauling the older parts of the WMATA Metrorail system is a plan known as Metro Forward:

The Metro Forward improvement program is an ambitious 6-year action plan that is designed to make your life better by making your commuting experience more reliable, more comfortable and more enjoyable.

We are working to provide you with new rail cars, new buses, new tracks, new technology and the rebuilding of essential infrastructure. Just over the coming year we will be rehabilitating and replacing nearly 50 escalators; renovating 12 Metrorail stations; retrofitting track and replacing track circuitry; rehabilitating third rail, running rail and track pads; installing track turnouts; purchasing new police radios; rehabilitating three bus garages; replacing 100 Metrobuses and rehabilitating 100 more. (http://www.wmata.com/about_metro/metro_forward/)

At a cost of $5.5 billion, the program started in 2011 is advertised to be replacing significant amounts of track, circuitry, escalators, and more. Focusing primarily on track for this article, Metro Forward is responsible for replacing on average 31,000 track fasteners and 18,000 ties per year, reducing the number of welds in the track, and completely replacing over 60 miles of track over the program's duration in addition to what is replaced during regular maintenance. (http://www.wmata.com/about_metro/board_of_directors/board_docs/100914_4DCombinedCOLOR.pdf)

Through October of 2014, WMATA shows that they have made some fairly significant progress in the program. Noted in the graphs below, Metrorail is working out of a backlog due to previous underinvestment in the rail system. One of the main items that is still unresolved and won't be for a few years is decreasing the number of track joints. Two joints are created each time a new piece of rail is inserted into the system - one at each end of the new piece - which end up causing a bumper ride for customers, and decreasing the lifespan of train wheels and the track itself. During maintenance windows, teams go around joining or welding these to the rail on either side in order to create one continuous piece of rail for trains to run on.

(http://www.wmata.com/about_metro/metro_forward/rebuilding.cfm)

As the charts show, WMATA is scheduled to make good headway into the backlog of work needed to restore the rails to a State of Good Repair (SOGR). However, what may be happening in the track/maintenance groups doesn't necessarily correlate directly with what customers see on a day-to-day basis. To see if it is, we turn to the public WMATA data.

Actuality

One of the documents published by WMATA is the Daily Service Report, a listing of all issues causing over ~3 minutes of delay to customers. All delays attributed to track problems are listed below. This may include anything from obstructions to cracked rails to smoldering rail ties. Unfortunately, the data from most of 2011 appears to be missing and not available on the WMATA website. In addition, some data has been misclassified either by WMATA or myself; the data used for this is as close to accurate as I have come so far. The period of time that we have data for encompasses the start of the Metro Forward program, which presumably included the most important work to get done first - including much of the track circuit replacement work required by the NTSB - but likely not the very beginning or the year directly leading up to its start. Due to this, it is hard to see what the state of the WMATA tracks was in 2011.

The data available so far include track problems starting in January of 2009 and continues through July of 2015 while omitting March 2011 through April 2012. There is more data to ingest pre-2009 as well.

WMATA Total Reported Track Delays per Quarter, 2009-2015

With 2011 data missing, we can clearly see the two sets of data from pre-2011 (approximately the start of Metro Forward), and data from 2012 to the present. Before the data gap, WMATA shows an average of 77 track problems per quarter between 1Q2009 and 1Q2011. After the gap, though, the average per quarter falls to 32 incidents. Unfortunately it is not possible to rule out that items are being classified differently over time as different individuals at WMATA file the reports over time, however it does appear to show a significant drop in reported track problems before and after Metro Forward started, or the general 2011 timeframe.

WMATA Reported Track Delays per Quarter, 2009-2015

When split up by rail line, this second graph is what we end up with. As shown the Red line is typically the main contributor to track problems. As the line is 32 miles in length by itself (of the total 117mi system length), this doesn't seem to be terribly unreasonable. In addition, track delays attributed to either the Blue, Silver, or Orange line through the core of the system technically can all count to the same amount of track for the portion which they share.

In both graphs, it appears there is a significant decrease in track delays in the pre-2011 data vs post-2011, presumably associated with the increase in track work started under Metro Forward. The main component of each set of data is still the Red line, however it also experienced a significant reduction in track problems since 2011.

There is a general ebb and flow to the data as well, which one would expect. Generally speaking there are more track problems in the winter and colder months (Q1 and Q2 covering January-June, primarily) compared to warmer ones. Track and rails are much more liable to crack and have defects in these conditions as the metal contracts. Not at all WMATA-specific, this condition plagues most/all railroads and is why some purposefully set their tracks on fire in order to keep running in the winter months. In addition when taking a look at the winter 2015 data, the harsher temperatures that we had are reflected in the number of track problems as a whole - in addition to the addition of the Silver Line entering operation.

Given this data, it does appear that significant progress in decreasing track problems happened during the 2011-2012 timeframe, early on in the Metro Forward program, and has resulted in a "new normal" of about 30 track problems per quarter. Likely, the "worst" spots were targeted first and fixed. Issues since then follow the cyclical weather cycle and do not appear to be rising significantly since (average 32, 24, 32, and 41 track problems per quarter in 2012, 2013, 2014, and 2015, respectively), though in a similar vein track problem occurrences do not appear to be decreasing either. At an average of about 33 track problems per quarter since 2011, a rider would expect 1 track issue throughout the system every 3 days.

While signal, switch, and track circuit issues were attempted to have been sub-classified separately, there are still a sizable number of track issues (although not apocalyptic levels). With additional work being done to reduce the work backlog to maintain a state of good repair, these numbers should stay relatively low and hopefully be reduced even more as time goes on. By all accounts, a track problem every three days is still too high and should be even further reduced as much as possible. This, however, is progress.


* Note: All charts and information based on the Daily Service Reports exclude March 2011 through April 2012 as there is a lack of usable data during this period.

# Note: Data cataloged by the Daily Service Reports is provided by WMATA and has been categorized to the best of my ability. The data here may contain errors and not be completely correct, although I have tried to minimize as many issues as possible. If you think something is in error, please feel free to contact me or leave a comment.