An undergraduate course was taught at the University of North Dakota (UND) that examined the objectives and motivation of the Deep Convective Clouds and Chemistry (DC3) field campaign. The course included discussions about particular challenges involved in making chemical transport measurements and field campaign forecasts; hands-on forecasting exercises for all three DC3 regions were also included. Subsequently, UND undergraduate students traveled to the University of Alabama in Huntsville to participate in DC3 campaign operations. While in the field, students actively participated in radiosonde operations (including instrument preparation and launch, field log documentation, and data quality control) as well as mission planning and daily forecast discussions. During the field period, a hands-on internship was held at UND that gave both graduate and undergraduate students a chance to forecast for the campaign (offline) and then listen in on the live forecast discussion. Formal assessment of the undergraduate course, the forecasting internship, and the campaign participation showed that students found all components valuable. Students gained experience in a variety of forecasting scenarios, forecast model evaluation, and field campaign decision making. This project represented a relatively inexpensive method to leverage the significant resources invested in field campaigns for greater educational gains. Recommendations are provided such that more colleges and universities can take advantage of opportunities to become educational partners with field campaign investigators in the future.
The University of North Dakota partnered with field campaign investigators to provide undergraduate students with an experiential learning opportunity that uniquely integrated classroom activities, operational forecasting, and a large multiagency field campaign.
It is crucial for the next generation of scientists, who will deal increasingly with research areas that cross not only disciplinary but also methodological boundaries, to gain as much cross-disciplinary and cross-methodological experience as possible. Large, agency-funded field campaigns represent this type of research, and as student involvement is normally a stated goal of these campaigns, providing additional avenues for formal undergraduate involvement is desirable. A successful methodology was recently tested and implemented at the University of North Dakota (UND) in partnership with the Deep Convective Clouds and Chemistry (DC3) campaign, the National Science Foundation, and the University of Alabama in Huntsville (UAH). Undergraduate students from UND participated in a focused spring semester course, a summer forecasting internship, and traveled to assist in radiosonde launches in support of the DC3 field campaign (Fig. 1). This project represented a relatively inexpensive method to leverage the significant resources invested in field campaigns for greater educational gains. Participation in field campaign activities, whether at the field site or remotely, is an invaluable student experience, providing motivation and encouragement. This participation can inspire students to careers ranging from private industry to academia and helps build potential collaborative relationships that can be the cornerstone of such careers.
Through the efforts of many universities and agencies, most field campaigns conducted have a valued outreach component, involving students at all levels as well as community members. These often include tours of the facilities for local student groups and recently have expanded into using social media (e.g., Twitter handle @DC3_Operations) to engage a broader student population as well as the general public. There have also been many successful efforts to involve undergraduate students directly in campaigns via educational deployments [see, e.g., EOL (2013a) for a list of recent campaigns]. Smaller deployments, like the Lake Thunderbird micronet project (Shapiro et al. 2009), have been very successful at integrating undergraduate students into field research and have the advantage of local, sustained deployment, leading to more hands-on time over multiple years. However, smaller campaigns cannot replace the experience of involvement in large deployments, including the opportunity to work with unique instrument platforms and meet researchers from national and international institutions. Such experience is particularly valuable for students at small, remote research universities and principally undergraduate institutions, where there is less frequent contact with the larger atmospheric sciences research community.
The integration of undergraduate students into large, multi-institution field campaigns can be challenging. One must create opportunities for undergraduate students to acquire the knowledge and skills that will allow them to be active and informed participants in campaign research, discussions, and other activities. While field experience for even introductory students has been shown to motivate and inspire, an understanding of the scientific objectives and measurement challenges of the campaign provides added value for upper-division students who are at critical decision points concerning their careers.
As one example, the Rain in Cumulus over the Ocean (RICO) campaign integrated 24 graduate and 9 undergraduate students into campaign operations (Rauber et al. 2007). RICO involved the students directly in both field measurements (using many different observation platforms) and operational forecasting. RICO also provided all the students with additional scientific background by running a seminar series at the field site during the campaign, which included research flights that were planned and executed entirely by students.
The goal of the UND project presented here was to provide a greater degree of integrated undergraduate participation. This integration took place by using three components to better engage the undergraduate students, specifically 1) a university course, 2) a forecasting internship, and 3) field experience. These components and assessments of each component are presented in the following sections. Additionally, the benefit to the students as future forecasters and researchers will be summarized in significant outcomes. Finally, recommendations for future implementations of this integrated approach will be given.
DEEP CONVECTIVE CLOUDS AND CHEMISTRY.
The Deep Convective Clouds and Chemistry (Barth et al. 2012) field campaign was conducted 15 May–31 June 2012. The primary goal of the campaign was to investigate the dynamical and chemical impacts of deep convection on the composition and chemistry of the upper troposphere. Several storm types were sampled, from airmass thunderstorms to mesoscale convective systems. The campaign involved in situ measurements from three research aircraft as well as many ground-based platforms, including radar networks, radiosondes, and lightning mapping arrays. Campaign observations were focused in three regions (northern Alabama, northeastern Colorado, and central Oklahoma), with flight operations based centrally to these areas in Salinas, Kansas. With such a broad sampling region, accurate forecasting of convective timing and convective type was crucial to campaign success. In addition, sampling teams were needed to staff the ground-based platforms full time in all three regions.
During spring 2011, a UND undergraduate course was taught that examined the objectives and motivation of the DC3 field campaign and included discussions about particular challenges involved in chemical transport measurements as well as hands-on forecasting exercises for all three DC3 regions. Based on student feedback from this initial offering, a spring 2012 course was offered allowing for more examples from the instructor and more hands-on learning in the forecasting laboratory with the aid of a graduate teaching assistant. Five undergraduate students participated in year 1 and six participated in year 2.
The two primary goals of the course were 1) to discuss scientific topics related to DC3 campaign objectives and 2) to provide experience in forecasting convective storms. The class started with an overview of the DC3 main objectives, with class discussions on various components of both the scientific program and experimental design. Specific topics that were covered in subsequent lectures included basics of atmospheric chemistry and transport, lightning mechanics and strategies for lightning observations, challenges of and strategies for a large and complex field campaign, and theory and forecasting of storm initiation and mode. Guest lecturers were invited to the class, including Dr. Larry Carey, the lead investigator from UAH (Carey et al. 2013).
The class forecasting experience was tailored specifically to the DC3 campaign by focusing on specialized regional convective forecasts using data from the previous spring's operations. In 2011, the course used data from the 2010 National Oceanic and Atmospheric Administration (NOAA) Hazardous Weather Testbed Spring Experiment (e.g., HWT 2013), and in 2012, the course used the 2011 DC3 test flights data catalog (EOL 2013b). The students developed criteria based on campaign goals and flight safety considerations that were deemed crucial for operations decisions: storm morphology, time of initiation, nearness to other storm complexes, and storm depth. The forecast validation involved specialized components such as examining storm depths with 3D reflectivity visualization. During the final week, the students produced forecasts for the beginning of the DC3 flight periods and were passive participants in the daily DC3 forecasts and flight-planning conference calls. We note that the course was helpful to the DC3 campaign scientists, as the students provided feedback to the investigators regarding specific observed and modeled variables that should be included in the DC3 campaign catalog during subsequent field phases.
Several assessments were conducted: 1) preassessment and postassessment of student confidence in abilities (e.g., different aspects of forecasting), 2) preassessment and postassessment of student confidence in understanding of DC3 campaign fundamentals (e.g., the impact of lightning on tropospheric chemistry), and 3) a UND university-wide standard course assessment. In addition, direct assessment of knowledge was conducted via 1) oral presentations on individual case studies each representing a weeklong forecast period and 2) multiple forecast and verification exercises for the three focus regions, including determination of the most probable region for a successful mission and verification of storm depth and extent with radar retrievals. Students were assessed on their ability to explain their reasoning for forecast and mission decisions and to provide insightful analysis of their forecasting skill via the verification task.
The UND standard course assessment data showed that response to the class was extremely favorable, with all students answering “agree” or “strongly agree” to “this course was a worthwhile addition to my university experience” (with 100% answering strongly agree for the 2012 course). All students also said the course encouraged making connections to real-world situations. The students appreciated the unique opportunity to discuss lightning physics, mesoscale and ensemble forecasting, and chemical transport. The live briefings were also appreciated by the students as an opportunity to hear how campaigns are run and how their own discussions might compare with those by the DC3 forecast team.
Table 1 shows questions asked in the pre- and postcourse student self-assessment of confidence in their own abilities and knowledge. Average scores from both the 2011 and 2012 courses are shown, with an additional column for a second postassessment conducted for the 2012 course approximately one year after the initial assessment. Comparison of pre- and postassessment of confidence levels showed some increase across all measured categories of forecasting ability, although increases were not dramatic as all but one student already had some forecasting experience (Table 1). Students' assessment generally reflected growth of their understanding of physical processes rather than growth in forecasting ability. Even higher scores characterized the assessment one year after the course ended; this trend is likely due to continued exposure to campaign goals while participating in the campaign as well as during preparation of American Meteorological Society meeting presentations related to their field experiences.
Increase in confidence does not necessarily correlate to a measurable increase in skill (Dunning et al. 2004; Sitzmann et al. 2010), as students tend to be overconfident in newly acquired skills and self-assessment is not indicative of long-term retention. However, for this course, instructor assessment of skill increases in forecasting ability and grasp of mission goals was proportional to the self-assessment results, although in some cases the self-assessment numbers were higher.
The course successfully utilized an experiential learning paradigm, providing students with experience in forecasting, field campaign decision making, and model evaluation. It was noted that the group discussion of the team forecasts and also next-day validation enhanced learning, but more benefit could be derived from more formalized self-assessments and reflection on problem-solving techniques. Others have noted the added value that methodical problem solving and self-assessment provides to undergraduate atmospheric sciences education (Bals-Elsholz and Goebbert 2013; Beyerlein et al. 2007). Thus, a future implementation could benefit from added reflection.
In summer 2012, coincident with the DC3 field campaign, six UND students participated in a forecasting internship associated with the campaign. The internship students provided daily forecasts directly related to DC3 campaign operations or, during quiet-weather periods, other forecast problems that were at least tangential to principal DC3 objectives. The UND forecasting team met 2 h daily, 7 days per week, for the 6-week duration of the DC3 campaign (with a few exceptions corresponding to consecutive DC3 hard down days) plus 3 days ahead of the official start of the campaign as a “warm-up” period. The forecasters began work on their forecasts 1 h prior to the scheduled DC3 daily weather/operations briefings, with the last 15 min devoted to student presentation and discussion of their forecasts. This discussion period was followed by approximately 1 h of passive participation in the DC3 briefings teleconference and a postcall debrief and question period.
The primary forecasting paradigm focused on convective probability from 0 to 36 h, using convective probability charts similar to those posted to the regional forecast summary under the daily reports section of the DC3 field catalog (EOL 2013c). During the internship, the daily forecast forms submitted by the student forecasters covered one or more DC3 regions and included convective timing, convective mode, and probability in 10% or 20% increments. The forecast paradigms were later modified to also include extended range (3–7 day) convective forecasts, chemical constituent forecasts (alone or in combination with convection), and prediction of Tropical Storm Debby (late June 2012). These modifications were made to provide opportunities to forecast for chemical constituents and lightning (as fitting with the DC3 objectives) as well as to add variety to the forecasts during quiet-weather periods. Submitted forecasts were accompanied by a multiparagraph discussion explaining the forecast rationale and relationships to expected DC3 operations. Initially, the students submitted their own individual forecasts, but team forecasts were gradually introduced in order to provide the students with some experience working in a team environment as might be found in the public or private sector.
Multiple types of assessment were done for the forecast internship, including a comprehensive postinternship student survey, examination of the progression of student forecasts over time, and instructor observations. Note that these results include all forecasting participants, both undergraduate (two) and graduate (four). The graduate students were in their first or second year of graduate school and so had similar responses to those of the undergraduate students. The survey asked the students to rank on a six-point scale (1 = awesome; 6 = terrible) their general likes and dislikes concerning the class and utility of specific aspects. In addition, the undergraduates were asked about how the spring university course prepared them for the internship and the value of the internship as a complement to field campaign participation. The internship was favorably reviewed by the students, with an overall score of 2.2. The highest-ranking components were the variety in forecast paradigms, the students' own improvement in forecast skill, and the value as a complement to field campaign participation. The lowest-ranking components were the early start time (0730 LT), the DC3 teleconferences, and forecasting for more than one region simultaneously. Based on their comments, the students had overly high expectations for the teleconferences, expecting exciting decision-making discussion, when much of the actual flight decisions are done offline and then updated once aircraft are in flight. The students also stated that they would have liked more competitive aspects (i.e., more team days), more discussion on interesting weather days, and more self-verification. The students also wanted more responsibility and the chance to be actual campaign forecasters.
Instructor observation indicated that the students were not well prepared for the pressured/multitasking forecast environment of the internship setting. As many students in their careers may sometime face this type of environment (e.g., private-sector forecasting), more exposure for students to such an environment is important. An important lesson learned from the internship was that there is a trade-off between variety and mastery. Although the students wanted and appreciated the variety provided by the modified forecast paradigms, this limited the amount of practice with any single paradigm. Such trade-offs do not generally exist in traditional forecasting curricula.
STUDENT FIELD EXPERIENCE.
The six undergraduate students from the spring 2012 DC3 course participated in the data collection in Alabama, going down to the field region for two weeks each in teams of two. The UND cohort took primary responsibility for the mobile soundings. A UND graduate student assistant was in Alabama for the entire campaign, leading the sounding team and helping the transition between different undergraduate teams. While in the field, students actively participated in all aspects of radiosonde operations, including instrument preparation and launch, field log documentation, and data quality control. They also participated in daily mission planning and forecast discussions. The students were given the opportunity to tour and learn about other observational platforms deployed at UAH, including radar and profiler platforms.
Feedback from the UAH team was extremely positive, commenting on the students' commitment and professionalism. In addition to being a valuable experience for the students, Dr. Carey commented that it would have been difficult to keep the multiple observational platforms fully staffed without UND's involvement. In one-on-one discussions during and after the campaign, students stated that they found the campaign participation to be very valuable. In particular, students enjoyed being able to relate classroom concepts to their field experience. They were excited to meet and work with “so many people from across the globe to work together on one common goal.” The students worked as a team after the campaign and put together an overview of their experiences for the American Meteorological Society student conference.
This project has been shown to be valuable by both increasing the undergraduate student participation in DC3 and providing multiple experiential learning opportunities for student participants. The specific benefits to the students are many, including increased confidence in forecasting ability, immersion in the science and implementation of a major field research campaign, and improved understanding of their own career options and interests.
Additionally, the benefits to student research ability can be even greater. One of the participants completed a research project as part of the forecasting internship and produced a thoughtful and useful study validating the campaign convective forecasts. The instructor noted exceptional gains in research ability over the course of the project. This same student is now using sounding, lightning mapping array, and radar data from the UAH DC3 observational platforms (in partnership with Dr. Carey) to research the relationship between lightning characteristics and convective regime.
As part of the 1-yr postassessment, students were asked to reflect on all three components of the project: university course, summer forecasting internship (if applicable), and field campaign. Not surprisingly, students chose the field campaign as their favorite component. However, reasons why participation in the field campaign was ranked as best varied and included seeing concepts from class being practiced in the field and being part of the Alabama forecast discussion, showing that the three class components tied closely together. Students also appreciated the large amount of responsibility that they were given as an independent observational team. Students were also asked for suggestions for future field campaign integration projects. Students suggested longer campaign involvement and possibly at more locations (although there was acknowledgment of funding constraints) and more time spent in class on the instrument platforms for which the students would be responsible (in this instance, more on radiosondes).
All the students commented that their project involvement impacted their approach to current coursework. Several students highlighted the impact of working with data directly, with one student stating, “I have more insight into thinking about how data was acquired, what are the possible errors with it, or what could have gone wrong while recording the data. [The project], in general, added another critical thinking mindset that ‘regular’ studies would not allow.” This ability to translate insights from their field experience into the classroom has the potential to improve their own classroom success (e.g., Etherton et al. 2011). Finally, students were asked if participation in this project impacted their career planning. Most of the students stated there was no impact, likely as several students had already planned to pursue graduate school and research. However, one student stated this experience did encourage “being more research-oriented in my endeavors in the future.”
RECOMMENDATIONS FOR FUTURE IMPLEMENTATIONS.
As shown in this study, student involvement in field campaigns is a relatively inexpensive but highly valuable activity, and integrated student involvement is particularly valuable. Below is a recommended strategy for involvement in future campaigns.
Identify candidate field campaign and partner with participants.
Major field campaigns are planned 1–3 years in advance of implementation, and it is often possible to learn of planned campaigns from government laboratories, program officers at funding agencies, and colleagues in the area of study that interests you. Once you have found a possible campaign to work with, contact the principal investigators (PIs) for further discussions as to whether the joint endeavor is a good fit. Educational partners can be extremely valuable to investigators, both providing broader education impacts and a pool of trained students to help with the personnel-intensive tasks of making observations, providing nowcasts/forecasts, and performing postcampaign quality assurance/control on the data. Discussions with campaign investigators and program managers should happen as early as possible to best integrate with the campaign objectives. Early discussions also provide adequate time to find the best funding model to fit the educational project proposed.
Plan course (and/or internship) and field experience.
Listed here are the key components that led to success of each component:
materials from both partner institution and campaign PIs;
focused lectures/discussion in subject areas specific to the campaign objectives, including special guest lecturers and field trips; and
experiential learning activities (forecasting, collecting or processing of observations, etc.).
forecasting exercises conducted in real time, covering a variety of paradigms and forecasting goals, primarily those tied to the field campaign;
access to products used and produced by campaign forecasters; the latter provides a starting point for discussion of forecast techniques, the forecast process, and modifications to that process in pursuit of specific campaign goals; and
active participation by student forecasters in campaign (not done for DC3, but this would enhance training for students and broaden the pool of forecasters).
active participation in daily mission planning and forecast meetings on-site, with adequate space to include all participants in order to enhance integration;
multiweek experience for each team of undergraduates; and
leader (e.g., graduate student, faculty member) to provide continuity between different undergraduate teams.
The authors thank Dr. Larry Carey from the University of Alabama in Huntsville for support of this project. The expertise and materials provided by Dr. Carey helped shape the course activities, and the training and facilities provided by Dr. Carey and his team at UAH were crucial in making the student field experience a success. The authors also thank all the UND students involved in this project for their enthusiasm and dedication. The authors would also like to thank the DC3 principal investigators, particularly Drs. Mary Barth and Chris Cantrell, for their strong support of this project and Alison Rockwell for providing DC3 outreach materials and including this project on the EOL DC3 blog. The authors appreciate the valuable feedback from Mr. Sean Arms and two anonymous reviewers. This research was supported by NSF EAGER Award AGS-1212279 and NSF Award ATM-0918010.
*CURRENT AFFILIATION: Division of Atmospheric Sciences, Desert Research Institute, Reno, Nevada