Fourth of five parts (see one, two or three)
Predator purchases ended in 2009, with a total of 248 being bought by the Air Force. Reaper purchases started in 2002, rose from four per year in 2004 to 48 per year in 2011, yielding a 108-strong Reaper fleet authorized by the end of 2010, with 48 more to be bought in both 2011 and 2012.
Previous plans for combined Predator and Reaper production had been to support 65 CAPs (four air vehicles each) by 2013. However, 2013 budget materials clarified that the 65 CAPs would not be complete until later, variously stated to be either 2014 or 2017. Air Force budget documents for 2013 assert that by the end of 2011 there were 60 Predator/Reaper CAPs, implying a total count of Predators and Reapers of 240. It seems like some of the Pentagon’s drones are missing…
Assuming a two year delay between purchase authorization and delivery, a combined total of 248 Predators and 108 Reapers (356 of both) should be available at the start of 2012. (If one assumes a one year delivery delay, that number would be 404 in 2012.) With almost all Predators and Reapers operationally deployed, there would seem an excess of up to 116 Predators and/or Reapers (or an excess of 164 if one assumes only a one year delivery delay.)
There are two potential explanations: either there are significantly more than four air vehicles per CAP (to address the infrequency with which they fly, an issue addressed below) or there have been an extraordinary number of Predator and Reaper crashes. The latter appears to be the more complete explanation.
Crashes: While it is conventional wisdom that drones are prone to crashes, that wisdom seems to understate the dimension of the problem.
As a gross indicator of the seriousness of the problem, DOD had expected the Reaper inventory to be 256 air vehicles in 2017, a year after the previously planned buy of 396 was to be complete. Assuming the last purchase is delivered within a year, losses as high as 140 air vehicles appear to have been anticipated. If the delivery lag is two years, not one, 92 air vehicles would seem to have been anticipated as losses.
The Air Force claims that it has reduced the loss rate for Predator from 28 mishaps per 100,000 hours to 7.6 and that Reaper can or will share this reduction in losses by virtue of its triple redundant flight controls, back up communications and other characteristics. The available data do not appear to support this claim.
While Air Force “mishap” reports show only three Reaper crashes since 2006; that data base is incomplete. A larger number has been publically reported, as well as an extremely high crash rate of 16.4 for every 100,000 flying hours.  A different public data base, at Drone Wars UK, appears to be a little more complete. It reported four Reaper crashes in 2011 and six more in the preceding three years. While the data show some evidence of a declining crash rate per flying hour as operators become more familiar with the Reaper’s characteristics, the data also show an increase in crashes, and the crash rate, from 2010 to 2011. However, the Drone Wars UK data base, being reliant on public reports, asserts “This list is almost certainly not complete.” The data here is clearly an undercount of Reaper crashes; an audit of each Reaper produced and its status and history is clearly needed to resolve this critical, possibly crippling, issue.
A high Predator crash rate tends to validate these data. CRS reported 20 Predator mishaps for every 100,000 flying hours in 2005. The Air Force claims that contemporary mishap rates have fallen as the system matures. Air Force class A mishap reports show 12 Predator MQ-1B crashes in 2011: a calculation of the crash rate per flying hour does not support the Air Force contention of lowering rates. Both the total number of Predator crashes and the rate per 100,000 hours of flying increased, for example, from 2010 to 2011. The Predator crash data at the UK Drone Wars website also shows a continuing and severe Predator crash problem. And, the gross numbers of Predators and Reapers produced compared to those operating in the inventory suggests that the number of Predator/Reaper crashes could be as many as 100 air vehicles, possibly more. Clearly, a full and independent audit of the fate of each tail number is called for to measure—definitively—the precise dimension of this problem.
Operational Availability: DOT&E found that Reaper has failed to meet its own criteria for system failures sufficiently seriously that it is not ready, several years after its operational deployment, for resumed operational testing. (The system continues to be deployed under an extended “Interim Authority to Operate.”) The system’s unreliability surely exacerbates the system’s high crash rate, and it is one of several factors impacting the system’s operational availability.
In 2011, the declared operational inventory of 69 Reaper air vehicles flew a total of 97,727 hours. That calculates to 1,416 hours per air vehicle per year, or 118 hours per month, or 29.5 hours per week. For a Reaper that flies the maximum 42 hour sortie (using two wing fuel tanks and two munitions), the air vehicle gets into the air less than once a week. If a Reaper flies what is described as a more typical 14 hour mission, it will be in the air twice a week, on average. Thus, individual Reaper air vehicles fly from less than once a week to as much as twice a week.
Some unofficial websites describe Predator and Reaper CAPs providing 24 hour per day seven days per week coverage of the battlefield. With four air vehicles, as defined by DOD, a Reaper CAP is incapable of providing that coverage: there are 168 hours in a week; any four Reapers flew, on average, 118 hours in any week in 2011. A six Reaper “CAP” could provide the 168 hours needed per week, providing a slim additional margin of nine hours, but if two hours of flying are required for transit to the operational area and back for each sortie, seven Reaper air vehicles would be required for a 24/7 CAP.
These sortie rates are a small fraction of what manned aircraft have flown in historically relevant combat. In Operation Desert Storm in 1991, GAO found that throughout the course of the 41 day air war, F-16s flew more than one sortie per day, not two sorties, or less, per week. The A-10 flew significantly more often than F-16s in Desert Storm. The notoriously unreliable F-111Fs flew almost one sortie per day (0.9), and the even more difficult to support F-117 had a rate of .7 sorties per day. In 2011, the average Reaper air vehicle flew somewhere between 0.1 and 0.3 sorties per day, a rate that for a single manned aircraft in sustained combat would surely be deemed catastrophically low.
An astute reader will observe that this analysis ignores Reaper’s much longer endurance (from up to 14 hours to as long as 42 hours) which enables it to search for targets and intelligence. To make the argument that Reaper’s (and other drones’) ability to loiter and search for and find targets proves superiority over manned aircraft in a critically important dimension assumes that the drones are effective at finding targets and collecting intelligence and at prosecuting the targets once detected and identified. The available evidence, discussed in Part 3, shows that, as a practical matter, Reapers (and other drones) are less effective than simple, even primitive, manned Cessnas at finding and identifying targets.
Next, and last, article: So where is the “revolution”?
Winslow T. Wheeler is the Director of the Straus Military Reform Project of the Center for Defense Information in Washington.
 P. 16, “Selected Acquisition Report (SAR) RCS: DD-A&T (Q&A) 823-271; MQ-1B UAS Predator, as of June 30, 2010.”
 Pp. 15 & 18, DOD Predator SAR.
 P. 5, DOD Reaper, SAR. See also p. 4 of DOD’s 2012-2041 Aircraft Procurement Plan at http://www.airforce-magazine.com/SiteCollectionDocuments/Reports/2011/May%202011/Day25/AircraftProctPlan2012-2041_052511.pdf.
 P. 58, United States Air Force, FY2013 Budget Overview, SAF/FMB February 2012 at http://www.saffm.hq.af.mil/shared/media/document/AFD-120209-052.pdf.
 P. 1-2 of Program Acquisition Costs by Weapon System, February 2012, Office of the Under Secretary of Defense (Comptroller) at http://comptroller.defense.gov/defbudget/fy2013/FY2013_Weapons.pdf.
 P. 57, United States Air Force, FY2013 Budget Overview, SAF/FMB February 2012 at http://www.saffm.hq.af.mil/shared/media/document/AFD-120209-052.pdf.
 That would be 248 Predators purchased through 2009, 108 Reapers purchased through 2010, and 48 more Reapers purchased in 2011.
 In 2011 only 10 Predators and 5 Reapers in the total available inventory of both were not operationally deployed, according to operating cost data provided by the Air Force and available from the author.
 There is also confusion about the total number of Reapers actually existing in 2011 according to Air Force records. Four different DOD reports give four different Reaper inventory counts varying from 63 to 74. See p. 4 of “Selected Acquisition Report (SAR) RCS: DD-A&T (Q&A)823-424; MQ-9 UAS Predator, as of June 30, 2010;” p. 5 of “Report on Future Unmanned Aerial Systems Training, Operations, and Sustainability,” Report to Congressional Committee, United States Air Force, September 2011; p. 21 of http://www.defenseinnovationmarketplace.mil/resources/UnmannedSystemsIntegratedRoadmapFY2011.pdf , and Air Force inventory, flying hour and cost data in the files of the author. Reapers are also being produced for the CIA; their numbers are classified and are publically unknown.
 See p. 5 of United States Air Force, Report to Congressional Committees, “Report on Future Unmanned Aerial Systems Training, Operations, and Sustainability,” September 2011, submitted to Congress pursuant to House Report 111-491, page 509.
 P. 32, CBO, “Policy Options,” at http://www.cbo.gov/ftpdocs/121xx/doc12163/06-08-UAS.pdf.
 See the Air Force’s official mishap data at http://usaf.aib.law.af.mil/index.html.
 For example, see “War Zone Drone Crashes Add Up,” David Zuccino, Los Angeles Times, July 6, 2010, at http://articles.latimes.com/2010/jul/06/world/la-fg-drone-crashes-20100706.
 See Drone Wars UK, Drone Crash Database, at http://dronewarsuk.wordpress.com/drone-crash-database/.
 P. 18, CRS, “U.S. Unmanned Aerial Systems,” at http://www.fas.org/sgp/crs/natsec/R42136.pdf.
 The Air Force claims that contemporary drone mishap rates have fallen to the point where they are comparable to aircraft, such as the F-16 at a similar (implied early) stage of maturity. The data available do not support that assertion. Air Force class A mishap reports show 12 Predator MQ-1B crashes in 2011 after almost 10 years of employment: Predator is not an “immature” system. While early F-16C mishap reports and flying hours are not available; in 2011 there were four F-16C class A mishaps. According to Air Force flying hour data, the two platforms flew approximately the same number of hours in 2011, just over 200,000. The Predator mishap (rather crash) rate calculates to three times that of the F-16C.
 P. 247, DOT&E 2011 Annual Report, at http://www.dote.osd.mil/pub/reports/FY2011/.
 The data is available in USAF flying hour and cost data made available to the author; these data are available on request.
 The 2011 data for Reaper’s “sortie” rate is not an exception. The 2010 data shows Reaper in the air for 30.4 hours each week, and the 2010 data for Predator (MQ-1B) shows it in the air 26.4 hours per week.
 For example, see “Predator Patrols to Nearly Double by 2010,” Defense Update, Undated, at http://defense-update.com/newscast/0707/news/200707_predator.htm.
P. 166, GAO, Evaluation of the Air Campaign,: at http://gao.gov/assets/230/224366.pdf.