Over the past two decades, the National Science Foundation’s Division of Atmospheric and Geospace Sciences (AGS) has funded nearly 200 atmospheric science–related field campaigns that have included deployment of AGS-sponsored observing facilities. These projects have spanned the range from modest, single-investigator experiments to massive, multi-investigator, multiagency campaigns. They have occurred both domestically and abroad, on every continent and over most oceans. In this article, we present an analysis of some of the details about these campaigns, including such elements as deployment location and cost of the campaign, and of statistics related to the principal investigators (e.g., type and location of institution, gender, years since degree). In addition, we assess trends in field campaign cost. These results provide a retrospective view of atmospheric science field work that has been supported since 1992.
Over the past two decades, programs in the National Science Foundation (NSF) Division of Atmospheric and Geospace Sciences (AGS) have funded nearly 200 atmospheric science-related field campaigns that have included deployment of AGS-supported observing facilities. These projects have spanned the range from modest, single-investigator experiments to massive, multi-investigator, multiagency campaigns. They have occurred both domestically and abroad, on every continent and over most oceans. In this article, we present an analysis of some aspects of these campaigns as well as statistics related to principal investigator demographics.
The data presented here were compiled from historical records maintained by the various facility managers that included dates and locations of field campaigns as well as information about the facilities deployed. Details of funding, both in support of the facilities used for the field campaign and the science associated with it, were obtained from NSF’s internal records. Only proposals funded for participation in a field activity were considered; proposals funded subsequent to a campaign (e.g., for data analysis or synthesis, modeling) were not included. Typically, these proposals were three-year awards that covered preparation for the field campaign, execution of the field phase, and initial data quality control and analysis. All costs have been normalized to 2009 dollars, based on Office of Management and Budget tables.1 Demographic data on principal investigators (gender, year of highest degree) were determined from information in the public domain, such as an individual’s professional web page, professional society sites, or LinkedIn.
OBSERVING FACILITIES AND DEPLOYMENTS.
The Lower Atmosphere Observing Facilities (LAOF) supported by NSF AGS between 1992 and 2015 are listed in Table 1. Note that few of these facilities were available for the entire period discussed in this article; some have been acquired or added, while others have been retired. This evolution of the deployable assets reflects both the natural life cycle of facilities and the changing needs of the observations community.
Since 1992, NSF AGS has supported, all or in part, 199 field projects. Complete funding data were available for all but seven of these campaigns. The vast majority of projects were funded by the Physical and Dynamic Meteorology (PDM) program or its predecessors (91), with an additional 30 supported by the Atmospheric Chemistry program. The remainder were distributed among Climate and Large-Scale Dynamics, Aeronomy, Solar-Terrestrial Physics, Ocean Sciences, Polar Programs, and various special initiatives.
The maps in Figs. 1 and 2 illustrate the deployment locations, domestic and international, respectively, associated with these field campaigns. Only the primary base of operations for each project is noted, not regions covered by aircraft flights or mobile ground-based campaigns. While deployment sites are reasonably well-distributed throughout the United States, there are some areas of concentration that reflect the predominant locations for certain types of phenomena under study. For example, the U.S. Central Plains is frequently chosen for experiments involving severe weather, while the California coast allows access to the marine environment. International deployments have been concentrated in the Western Hemisphere, with a smattering in Europe and Asia. Generally speaking, this is related to ease and safety of deployment, rather than about the presence or lack of interesting phenomena to study. NSF-funded field campaigns in Africa are notably lacking. Although several have been proposed in recent years, logistical complexity, high cost, and safety concerns have prevented these projects from being selected.
FIELD CAMPAIGN COSTS.
Figure 3 (left) shows the combined total cost of each funded field campaign since fiscal year 1992. These values include the costs of both deploying the supported facilities and those that are directly associated with the science of the field campaign, borne by the science programs. All values have been normalized to 2009 dollars, as noted above. These same data are shown as a histogram in Fig. 3 (right) to illustrate that the majority of campaigns (∼62%) cost less than one million dollars ($1M). Nonetheless, slightly more than 10% of funded campaigns in this time period cost in excess of $5M and 1.5% exceeded $10M. The most expensive study carried out in the 1992–2015 time frame was Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE); the remainder of the “top 10 list” is presented in Table 2.
There is a sense among NSF program officers and LAOF facility managers that atmospheric science field campaigns have grown more complex and more expensive over time. While it is difficult to objectively assess “complexity,” which can include factors such as number of participants or facilities, deployment location, number of funding agencies involved, number of scientific objectives, etc., the normalized cost data do allow us to look at the expense of projects. The median cost of field campaigns for each fiscal year was computed and is plotted in Fig. 4, along with a linear fit line (r2 = 0.34) that indicates growth in cost of approximately $61,000 per year, resulting in nearly a quintupling of the median cost over 20 years. The high median costs in a few years skew this trend, but they do not necessarily correlate with years in which the most expensive field campaigns were supported; rather, there is an absence of low-cost (<$100,000) studies. Removing the values for the five highest-cost years results in a trend (r2 = 0.32) of $25,000 per year. By comparison, a linear fit to the annual mean of field campaign funding over time (not shown here) yields an increase of about $62,500 per year (r2 = 0.25), similar to the value obtained from the median. This suggests that occasional very expensive campaigns are not making a large contribution to the calculated trend. Nonetheless, the interannual variability is quite large, so any conclusions about the magnitude of the increased cost of field campaigns over time are difficult to draw.
Demographic data were compiled for the people listed as principal investigators (PIs) on each of the 199 field projects. Most projects had more than one PI or lead scientist, identified through their coordination of the lead scientific proposal, science planning document, or lead author of the main campaign publication. PIs of individual grant proposals associated with projects were not enumerated. Among the 327 PIs, there were 190 unique investigators; of these, only 24 were female. It was not possible to reliably determine race or ethnicity based on the data available, so minority status is not reported here. PIs were distributed across 70 different institutions or organizations. Not surprisingly, the majority (51) of these were doctoral institutions (40 R1 and 11 R2).2 The remainder came from four master’s schools (3 M1 and 1 M3), three baccalaureate colleges, seven small businesses, one Federally Funded Research and Development Center (FFRDC), two U.S. Government laboratories, one nonprofit, and one four-year engineering-focus school. However, more than half of PIs represent just 11 institutions: National Center for Atmospheric Research (NCAR, 42 PIs); University of Colorado (10); University of Washington (9); University of Wyoming (8); University of Oklahoma (7); Colorado State University (6); University of Utah (6); University of Nevada, Reno/Desert Research Institute (5); The Pennsylvania State University (5); University of Illinois (5); and University of Miami (5). The western U.S. university focus is curious, but we do not know the reason for it. The presence of three facility provider institutions in this list (NCAR, University of Wyoming, and Colorado State University) reflects both the leading role that scientists from those organizations play in their fields and also the funding of numerous test or small exploratory campaigns originating from facility-affiliated scientists.
A long-standing concern in the atmospheric and related sciences that there is a shortage of researchers trained in observational science prompted us to look at the experience level of scientists proposing successful field campaigns, based on years since highest degree (generally the Ph.D.). Figure 5 shows a histogram of the distribution of PI experience, binned in 5-yr increments. Only about 16% of PIs can be considered early career (≤10 years since degree), while nearly 20% have more than 30 years of experience since receiving their degrees. The relatively small number of early career PIs is not surprising, given that leadership of field campaigns can be a high-risk activity that is discouraged in the pre-tenure years. If, instead, the average PI experience (that is, the average of the experience of all PIs on a campaign) is distributed, the shape of the histogram is quite similar, with 12% of campaigns having a PI team with an average of 10 years’ experience or less. This suggests that partnering of junior PIs with more senior ones is not very common, even though that might be a way to reduce the risk to an early career scientist and for less experienced researchers to gain valuable mentoring.
In Fig. 6, the data from Fig. 5 are broken down by gender. The distribution of experience among the female PIs is much more compact than that of male PIs, most likely reflecting the more recent entry of substantial numbers of women into the atmospheric sciences (see data in Hartten and LeMone 2010). Interestingly, though, substantially higher percentages of younger women than men have successfully led field campaigns in the past two decades.
Finally, we look at whether there is a trend in PI experience with time. Figure 7 shows the average years since degree of PIs for field campaigns conducted since 1993. Although the correlation is not terribly strong (r2 = 0.58), there does appear to be a trend of increasing years of experience with time. This might be attributed to the idea that, once successful, PIs successfully propose further in later years. However, this increasing experience with time runs somewhat counter to expectations, given that more, collectively less experienced, women are entering the field. Recalling the trend toward more expensive (and perhaps more complex) field campaigns, we note that there is no correlation between campaign cost and PI experience (not shown). While it is hard to know what the root cause of this trend is, it is worrisome that the population of field campaign PIs is aging.
In this short article, we have presented an analysis of some aspects of the past twenty years of NSF AGS-supported atmospheric science field campaigns to illustrate both general characteristics and possible trends. We have not attempted to quantify the results of these field campaigns in terms of publications or other outcomes; while interesting and valuable, that type of analysis is beyond our current data-gathering capabilities. It is important for the reader, and especially for anyone who might be planning to propose a field campaign, to realize that nothing presented here is intended as a guideline or constraint. Rather, we encourage early and direct communication with NSF program managers and LAOF providers to obtain advice about what is feasible. In particular, we note that many successful field campaigns have been funded only after revision and resubmission, building on feedback from the NSF merit review process and the facility request review conducted by the Observing Facilities Assessment Panel (OFAP). We anticipate that the creative and enterprising atmospheric sciences community will continue to execute exciting field campaigns well into the future.
Disclaimer: Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
FOR FURTHER READING
Table 10.1, available at https://obamawhitehouse.archives.gov/sites/default/files/omb/budget/fy2017/assets/hist10z1.xls.
According to the Carnegie Classification of Institutions of Higher Education—see http://carnegieclassifications.iu.edu/classification_descriptions/basic.php.