For the last five years, the author has employed a business process model as a central organizing construct for the senior-level Forecasting Techniques course at Embry- Riddle Aeronautical University's Daytona Beach, Florida, campus. The process model allows weather analysis and forecasting to be examined as both a scientific process and a business operation, with emphasis on employing a user-focused approach. The use of the model arose from the need for an organizing context for the students, mostly seniors applying their knowledge from previous coursework, most of whom are learning to make their first weather forecasts. The process model used in the present version of the course evolved from one originally developed by the U.S. Air Force to describe weather information's value-added contributions to daily operations. The model consists of two major interrelated components: the weather information processing cycle (WIPC) and the provider–user relationship (PUR). The WIPC describes the analysis/forecast process from the scientific point of view, whereas the PUR examines the relationship between the provider and user of meteorological information. The WIPC uses familiar concepts such as data collection, analysis, and prediction, whereas the PUR introduces the students to complex (and seldom taught) topics such as user requirements and mission analyses. The process model also provides a framework for the final project, a case-study analysis that emphasizes not only the weather associated with the event but also its resulting impact on the affected population.
Seniors at Embry-Riddle University—many of them making their first weather forecasts—learn to see analysis and forecasting as both a scientific process and a business operation.
For the last five years, a business process model has been used as a central organizing construct for the senior-level Forecasting Techniques (WX 427) course at Embry-Riddle Aeronautical University's Daytona Beach, Florida campus. The Applied Meteorology Program has been granting undergraduate degrees since 2001, so it is a relatively young program with approximately 110 undergraduate majors and 180 minors. Embry- Riddle's undergraduate program offers five areas of concentration in aviation, media, commercial applications, computer applications, and research. Forecasting Techniques is a three-credit-hour course. It is normally taken second or third in a required four-course sequence that begins with Synoptic Meteorology (WX 356, an introduction to synoptic meteorology and computer applications). Forecasting Techniques can be taken either in conjunction with or after Advanced Weather Analysis (WX 456, a course that blends concepts from synoptic and dynamic meteorology and provides an introduction to mesoscale meteorology). The sequence ends with a capstone course, Weather Operations Seminar (WX 457), which introduces students to simulated and real-world forecast operational environments representative of various career paths that they may take upon graduation.
There are several motivations behind the business process model approach to teaching WX 427. The first is to provide a central organizing concept for the course. Because WX 427 is normally taken between WX 356 and WX 457, the students typically enrolled in the class are seniors, the majority of whom are beginning to apply their knowledge from previous coursework as they learn to make their first weather forecasts. At this stage of their education, it is important that they develop the proper “mental models” of analysis and prediction, especially as various concepts start coming together in the forecasting process; the process model provides a template to facilitate this.
A second reason for employing a process model in this course is to prepare the students for WX 457, which goes beyond basic forecasting skills and often has students working on team projects that incorporate the “business operations” portion of weather analysis and forecasting. The process model used in WX 427 contains two major components: one focused on the technical aspects of weather analysis and forecasting and the other on user-focused business operations.
A third reason for using the process model is to provide an organizing framework for the students' final project, which is based on detailed analysis of a historical event that was influenced by the weather or climate or a historical weather event. The students are required to employ both portions of the concept model in their case-study analyses in order to understand the weather event and its impact on the affected region (e.g., population, infrastructure).
This paper provides a brief background on the use of process models to describe the weather forecasting enterprise, a history of the process model used in WX 427, and a description of how the model is applied in the course. The paper concludes with a brief assessment of the educational methodology employed in the course.
USING PROCESS MODELS TO DESCRIBE WEATHER ANALYSIS AND FORECASTING.
According to Aguilar-Savén (2004), a business process is a combination of a set of activities within a business that describes the logical order of its activities and their dependence on one another. Business process modeling is a representation of those activities, which enables a common understanding and analysis of a business's key processes, deficiencies, and areas for process improvement. Aguilar-Savén reviewed a dozen methodologies for business process modeling, ranging from flowcharts to very structured techniques that can be translated into computer programs. Because industrial meteorology1 often includes weather analysis and forecasting operations, it is not a great leap to adapt business process modeling techniques in order to improve our understanding of those operations.
Most of the business process models used to describe weather analysis and forecasting have been of the flowchart type. In fact, these models have been quite useful to illustrate the evolution of the forecasting process itself. Dutton (2002) employed two flowchart-type diagrams (his Figs. 4 and 5) to compare traditional approaches for incorporating weather information in user decision making with an emerging approach where standard meteorological information produced by the federal forecast centers is converted by private-sector firms into impact information and decision aids for integration into user decision-making processes. Dutton's Fig. 5 is reproduced here as Fig. 1 for convenience of illustration.
More recent examinations of the forecasting process have focused on the user of the information in an attempt to improve our understanding of how weather information is employed in decision making and how different types of information are employed by different user communities. Morss et al. (2008) proposed a flowchart-based process model as part of the societal and economic research and applications portion of the North American The Observing System Research and Predictability Experiment (THORPEX) program. Their model examined the dissemination of weather information (including its uncertainty), the use of decision support tools by users, and how the users' outcomes could be fed back to weather information providers through a mixture of traditional meteorological forecast verification measures and value added/relevance to the user (see their Fig. 1). A related THORPEX concept paper on communicating uncertainty by Brooks and O'Hair (2006) also employed a flowchart-based process model. Their model described the production of traditional meteorological information [e.g., numerical weather prediction (NWP) model forecasts]; interpretation by forecasters (including the forecast uncertainty); and the outcomes of that communication in terms of “economic value, personal safety, and trust” (see their Fig. 1). In both of these papers, the user's interpretation of translated forecast information and the outcomes of that interpretation were very important components of the process models. In the WX 427 process model, concepts such as dissemination, integration into decision making, and the provider–user relationship (PUR) are used to convey ideas that are similar to these studies, albeit in less detail.
The process model used in WX 427 originated with the U.S. Air Force (USAF). H. Massie et al. (1995, unpublished manuscript) described the processes of data collection, analysis, and forecasting; applications to warfighter models; and dissemination in a concept paper that advocated investments in remote sensing technologies that could provide observations in data-sparse areas. They argued that the observations obtained from these remote sensing platforms (data collection) would form the critical foundation for all the process actions to follow, such as improved regional-scale NWP model forecasts (analysis and forecasting). The output from the NWP model would be used to develop user-focused applications for predicting parameters such as cloud ceilings, surface visibilities, and flight weather hazards. This information would then be fed to weather effects decision aids (applications to warfighter models) that would identify the potential impacts of the predicted weather on both friendly and enemy forces. The resulting impacts information would be made available to all levels of military operations through the development of a robust global communications network of military and commercial systems (dissemination). This approach to depicting weather analysis and forecasting as a continuous process allowed USAF weather leadership to provide more quantitative estimates of weather information's value added to daily operations. Lanicci (1998) subsequently captured this sequential, one-way process graphically using a flowchart-based process model (shown in Fig. 2). The evolution of USAF weather analysis and forecasting operations into a more continuously updating process that included information integration into the user decision-making process was described by Lanicci (2003) and is reproduced here as Fig. 3. Note that these military-focused papers highlighted the importance of developing accurate and relevant user-focused decision tools in order to optimize their use of weather information, not unlike the process models discussed by Dutton (2002) and Morss et al. (2008).
The description of the business aspects of weather analysis and forecasting is not restricted to process modeling approaches. A number of university programs have used several different techniques to incorporate the business operations portion of weather analysis and forecasting into their curricula, either directly or indirectly. Several examples of the differing approaches are presented here and are not intended to be all-inclusive. One method is a student-based forecasting- type operation in association with the department. Examples here include Penn State University's Campus Weather Service (http://cws.met.psu.edu/), the University of South Alabama's Coastal Weather Research Center (www.southalabama.edu/cwrc/index1.html), and the University of Wisconsin— Milwaukee's Innovative Weather (Roebber et al. 2010; http://innovativeweather.weebly.com/). Another method is to offer an “industrial meteorology” or “business/commercial” track as an option in the undergraduate curriculum, such as the University of South Alabama's Industrial Meteorology track and Embry-Riddle's Commercial Weather area of concentration. In these types of programs, students take courses outside the department, such as Marketing, Management, or specialized courses offered within the department. A third approach is unique and has been used by the University of Oklahoma for the last 14 years, known as the Master of Science in Professional Meteorology option (MSPM; see http://som.ou.edu/degrees.php?program=mspm). The program is geared toward students desiring a private-sector career and includes two distinctive aspects: 1) requirement for 12 credits of study in a secondary area and 2) student sponsorship/support by a private company, which includes an applied research project, normally chosen and supervised by the sponsor (Carr et al. 2002; Carr 2008).
APPLICATION OF THE BUSINESS PROCESS MODEL IN WX 427.
The current version of the business process model employed in WX 427 is shown in Fig. 4. The model contains two primary but interrelated “tiers”: 1) the weather information processing cycle (WIPC) and 2) the PUR. The process model expands several of the modules from the earlier USAF versions and adapts the symbology of business process modeling notation (BPMN) described by White (2004).
The WIPC is shown as a rectangle, meaning that it is “owned” by “process participants” (BPMN nomenclature). These process participants are primarily public-sector entities such as the National Centers for Environmental Prediction (NCEP) and its product centers for the modules through the tailored products phase. However, as in Dutton's model, there is also participation by private-sector interests in the tailoring, dissemination, and user integration phases. The process participants are not shown as separate entities as required by a strict interpretation of BPMN but are shown together as a “community of providers.” This simplification is made so that topics such as division of responsibilities between publicand private-sector weather information providers can be introduced to students using the WIPC wit h minimal distraction from the symbology, while also illustrating that the public and private sectors can also partner. The individual process modules within the WIPC are shown by rounded rectangles, which denote activities (BPMN nomenclature), with the solid arrows denoting sequence f lows (BPMN nomenclature) as data move through the production cycle and are transformed into useable information.
In the PUR portion of the model, the provider and user are shown as process participants (in this case using separate rectangles to make distinctions between providers and users of meteorological information). The dashed arrows from user to provider and provider to WIPC process owner(s) denote message flows (BPMN nomenclature) between these groups, similar to the feedback processes proposed by Morss et al. (2008).
The process model is introduced to the students in multiple stages, beginning with collection, analysis, and prediction from the WIPC; transitioning to the PUR; and coming back to the WIPC modules of product tailoring, dissemination, and integration at the end of the course.
Collection, analysis, and prediction.
The first three WIPC modules, shown in Fig. 3 as a cycle using the solid arrows, denote the traditional meteorological data production cycle operated at many centers around the world. The author spends the first three weeks of the course on this part of the cycle for two main reasons: 1) to familiarize the students with the data from the various observing platforms they will be using in making and verifying their forecasts and 2) to introduce the students to the structure and operation of NWP modeling systems run at national centers such as NCEP. To support the WIPC lectures, homework questions are developed from assigned readings taken from Persson (2007, hereafter P07), which includes sections on NWP model equations, physical processes, data assimilation, and forecast verification and is written at a level quite understandable by senior undergraduates. Of particular interest is the forecast verification section of P07, which discusses both traditional verification scores such as mean absolute error and more forecast value-added measures such as cost/loss ratio. The forecast verification section is utilized extensively in the PUR, tailoring, dissemination, and integration portions of the course.
By modularizing portions of the WIPC, the author is able to address various aspects of the analysis and forecasting operation and introduce students to some of the issues associated with them. For example, in the collection lectures, we discuss the importance of satellite-based soundings in improving the analysis and forecast quality of NWP models by using comparisons of data-coverage maps from different collection platforms obtained from the Met Office's four-dimensional variational data assimilation (4DVAR) scheme (for details, see http://research.metoffice.gov.uk/research/nwp/observations/data_coverage/index.html). The plots are shown to illustrate the uneven spatial distribution of rawinsonde, aircraft, and surface data observations compared to those from the Advanced Television and Infrared Observation Satellite (TIROS) Operational Vertical Sounder (ATOVS) and the sheer quantity of observations available from satellite soundings compared to those from conventional sources. An accompanying 500-hPa height anomaly correlation time series from the European Centre for Medium- Range Weather Forecasts (ECMWF) NWP system, taken from Fig. 3 of Hollingsworth et al. (2002), is then shown to illustrate how the introduction of satellite-based soundings has closed the verification gap between the hemispheres and helped improve the quality of model forecasts in general.
The preceding collection discussion is a good segue to the analysis section, which employs the COMET module on data assimilation (available online at www.meted.ucar.edu/nwp/model_dataassimilation/) as a homework assignment and lecture tool in addition to P07, in order to introduce the students to the sophistication of today's NWP analysis systems. Although the intent of the analysis module is to give the students an appreciation for the complexity of data assimilation, it is just as important to give them some tools to evaluate the quality of model analyses and short-range forecasts. These quality check tools along with some suggested sources are listed in Table 1.
The prediction module takes the longest to complete. It is divided into multiple lectures along with a homework assignment on NWP. The lectures cover topics such as the history of NWP, model dynamics and physics, model postprocessing to include an introduction to model output statistics (MOS), and an intercomparison of operational models running at national centers such as NCEP. The lectures and homework are designed to give the students a qualitative appreciation for the complexity of the prediction model equations, physics parameterizations, and reasons for postprocessing raw model output. They also get introduced to the types of physical and dynamic features that a model can resolve as a function of its horizontal grid spacing and distribution of levels in its vertical coordinate system. One way to convey these differences in resolution is through comparing model depictions of topography in regions of complex terrain. An example comparing the Global Forecast System (GFS), North American Mesoscale model (NAM), and actual topography in Alaska is used in lecture because Anchorage International Airport (PANC) is one of the cities for which the students make forecasts.
The prediction module concludes with an introduction to forecasting techniques such as persistence/ modified persistence (important in Florida, especially during the wet season), climatology, analogues, and the forecast funnel (Snellman 1982). It is emphasized that an operational forecast of ten combines features of these techniques. It is also at this stage where the students begin making their first set of weather forecasts, a series of five city pairs for Daytona Beach and a second loc at ion out of the conterminous United States (CONUS). The list of cities and forecast parameters is shown in Table 2. These city-pair forecasts give the students an opportunity to put into practice what they learned in the first three WIPC module lectures and homework. For example, in the collection lectures, students are introduced to the operation of the Automated Surface Observing System (ASOS) and examine the observing procedures from OFCM (2005), particularly with respect to the FMH visibility reporting procedures. This part of the FMH becomes particularly relevant in the “WX = Y/N” portion of the exercises described in Table 2. The formulation of this forecast criterion forces the students to pay close attention to the observational data and not become totally reliant on the NWP products and MOS guidance. Additionally, the WX = Y/N parameter is used to illustrate the difficulties associated with making a “binary” forecast decision, which often frustrates both students and instructor alike when the forecast does not verify. The author subsequently builds upon this binary forecast experience in the second part of the course, when the students are introduced to making probabilistic-type forecasts.
The purpose of the PUR is to show that the relationship between providers and users of meteorological information has certain aspects that both sides need to understand and that this relationship can in turn affect the execution of the WIPC through user requirements (e.g., new observational data needs) and product feedback such as forecast verification statistics and value-added performance metrics. Within the PUR, we examine concepts pertaining to this relationship from the points of view of both the provider and user. These topics, summarized in Table 3, include several that are not traditionally covered in introductory forecasting courses, such as user-requirements analysis. It has been the author's experience that proper development of user requirements is a very complex process for which many scientific professionals are unprepared when they move into technical management positions later in their careers.
At this point, it would be useful to explain why the PUR is covered before the product tailoring, dissemination, and integration modules in the WIPC. There are three reasons: 1) Establishing the types of tailored weather information products needed by users is determined by analysis of the user's requirements. 2) The method of weather product dissemination is also largely determined by the user's requirements. 3) Integration of the weather information into the user's decision-making process is usually part of the user's business practices. In order to understand the latter part of the WIPC then, it is better to cover the PUR first. This portion of the course also marks the transition from the city-pair forecasts to a second set of forecasting exercises that reinforces the concepts taught in this phase of the course.
The PUR begins with a series of lectures on the types of meteorological knowledge that a provider needs in order to understand fully and satisfy the requirements of the user. This includes topics such as scales of motion; applications of descriptive, dynamic, and applied climatology; and knowledge of local effects/phenomenology. In-class examples are often used to cover several of these areas simultaneously. A good illustration is the Florida sea-breeze circulation. In the WIPC, the sea-breeze phenomenon was used as an example of the limitations in NWP model reproduction of mesoscale features, incorporating portions of the COMET sea-breeze module (available online at www.meted.ucar.edu/mesoprim/seabreez/index.htm). In the PUR, the Florida sea-breeze circulation is used to examine the interaction between processes operating on the synoptic scale and mesoscale in the scales-of-motion lecture but also as an example of local effects and phenomenology knowledge that a forecaster must have. The sea-breeze phenomenon is a very useful teaching tool in the course because the lecture materials are reinforced through forecasts for Daytona Beach International Airport (KDAB) and the National Weather Service (NWS) Melbourne forecast office's area of responsibility in east-central Florida.
The PUR discussion continues with an examination of the relationship between providers and users. Here, the author asks the question, “What does the provider need to know about the user?” This leads to discussion of the types of users of meteorological information, weather “salience” of the user (see Stewart 2009), determining how the weather and/or climate impact the user's business operation and how to determine user requirements and convey the provider's limitations to the user. In class, the author gives students examples of vague and specific user requirements, requirements that sound more like “solutions,” and unrealistic requirements (i.e., going beyond the provider's capabilities) and asks the students to distinguish among them.
The “flip side” of the PUR is examined with the question, “What does the user need to know about the provider?” In this section, the class covers points about determining the meteorological knowledge of users, how much users can articulate about how weather and/or climate impact their business operations (i.e., whether they have quantitative impact data or anecdotal experiences only), and how users may convey their requirements for weather and/or climate services to the provider (e.g., “I do not know what I need. What do you have?”). The author emphasizes the importance of ensuring the user understands the provider's capabilities and limitations, especially during the requirements analysis portion of developing their business relationship.
The PUR section closes with an illustration of one method for determining user requirements, by categorizing his/her “mission areas.” This essentially involves examining the user's weather and/or climate sensitivities and needs in three domains: 1) resource protection; 2) risk mitigation; and 3) exploitation. Resource protection is defined as the traditional mission of protecting life and property. Risk mitigation is defined here as “sustained action taken to reduce or eliminate long-term risk to people and their property from hazards and their effects” [taken from the Federal Insurance and Mitigation Administration, a division of the Federal Emergency Management Agency (FEMA); see www.fema.gov/about/divisions/mitigation.shtm]. Exploitation is presented as an emerging area of support to business and military users. It is distinguished from risk mitigation by examining the user's weather and/or climate sensitivities vis-à-vis a competitor, in order to determine if there are weather and/or climatic situations that the user can exploit to his/her advantage, at the expense of the competition. This portion of the course uses several hypothetical as well as historical examples to illustrate these ideas and emphasizes that there is nearly always a significant overlap among the three mission areas.
Product tailoring, dissemination, and integration into user decision making.
The PUR and WIPC are linked when we examine how different types of users (e.g., general public, businesses, aviation, and the military) employ tailored weather information and sophisticated dissemination technologies to integrate the information into their decision-making processes. At this stage of the course, the students are introduced to a new set of real-time forecasting exercises. These exercises give the students, now working in teams, experience preparing different types of tailored weather forecasts, varying from synoptic-scale products similar to those of the Hydrometeorological Prediction Center (HPC) to local forecasts for a hypothetical weather-sensitive customer. These exercises are summarized in Table 4. Although only the fourth type of forecasting exercise is totally linked to the PUR, the introduction of probability-based forecasts in exercise types 2 and 3 gives the students an appreciation of how a user would employ this type of information as opposed to the binary yes/no employed in the city pairs. At the completion of these last four exercises, the students are now prepared for the types of team-based projects they will execute in WX 457.
To drive home the point about being integrated into the user's decision-making process, the class goes on a field trip to the USAF's 45th Weather Squadron operations center at Cape Canaveral Air Force Station. This visit consists of a mission briefing; tours of the Applied Meteorology Unit and operations floor, complete with product demonstrations; and a briefing on how to develop tailored products that are timely, relevant, and useable for real-time decision making. The trip allows the students to get a firsthand look at how tailored weather decision guidance is integrated into the decision-making process for space launches.
ASSESSMENT OF THE EDUCATIONAL METHODOLOGY EMPLOYED IN WX 427.
The breakdown of grading in WX 427 has historically been a 50–50 split between forecasting exercises and homework/final project. This is done because the course is a combination of traditional lecture, homework, and forecast practicum, and the intent is for the students to learn about the forecasting enterprise by applying the WIPC/PUR business process model throughout the semester. In order to assess the effectiveness of the WIPC/PUR model as a teaching tool, the author used several methods. First, the author examined how well the students applied the WIPC/PUR model in their final projects by looking at the project grading distribution over the past eight semesters.2 Within this sample, the author also compared the grades for the four most commonly chosen project topics and looked for similarities and differences between them and the rest of the project topics. Second, the author investigated the results of end-of-course evaluations, paying particular attention to student ratings on course organization, clarity of objectives, and the overall evaluation of the course, these being a reflection of using the WIPC/PUR model as an organizing construct.
The final project consists of a written paper and an oral presentation, in which the students are instructed to apply the principles of the WIPC and PUR that they learned throughout the course. Specifically, they are asked to describe which portions of the WIPC were most relevant to their case and how well they worked and to evaluate the PUR in terms of areas such as provider knowledge of the user and user weather knowledge and awareness. Specific examples from historical events are used to illustrate aspects of the WIPC and PUR as a means of providing project guidance. For example, the WIPC processes of collection, analysis, prediction, and integration were most important for the D-Day invasion because the Allies' superior observational data availability over the northeast Atlantic Ocean led to a better analysis and forecast on the Allied side versus the German side. This forecast was effectively integrated into planning and execution to achieve the surprise necessary for the invasion's success. A standard PUR example used in class is the relationship among forecasters, engineers, and National Aeronautics and Space Administration (NASA) decision makers in the Space Shuttle Challenger accident. We discuss the lack of communication and improper integration of weather into the decision-making process that led to the tragedy (among several lessons that are reinforced during the 45th Weather Squadron field trip). When the students apply the WIPC/PUR model properly to their case studies, this is reflected in their project grades.
The final-project grades from spring semester 2007 through fall semester 2010 are shown in a box-and-whisker timeline plot (Fig. 5). The majority of project scores typically ranged from the upper 70s to the lower 90s. The mean grade over this period was 85.5 with a standard deviation of 7.4; there were 51 total projects encompassing 109 students. This result suggests that the majority of students were achieving a reasonable understanding of the WIPC/ PUR model as it applied to their final-project cases. However, Fig. 5 also shows an increase in the range of grades after fall semester 2008, with an increase in the number of project grades below 70 during the last three semesters. The project guidance remained relatively unchanged throughout this period with one notable exception. Beginning with fall semester 2008, the term paper due date was changed to three weeks earlier than in previous semesters. This change was made to give students the opportunity to make revisions prior to the end of the semester; the presentation dates remained unchanged. The rationale was that the review and revision process would provide a valuable learning experience, potentially leading to a final written product suitable as a writing sample for prospective employers. However, it is quite possible that the earlier due date resulted in some students having an incomplete grasp of the WIPC and PUR, with the revision process becoming more of a “dot the i's and cross the t's” exercise than a true manuscript revision. The mean project grade/standard deviation before the due-date change was 86.3/5.9; after the change it was 85.1/8.2. The author ran a z test on the grading data from before and after the due-date change; the test results showed that the means did not differ significantly. However, it is entirely plausible that an inability to grasp the WIPC/PUR concept fully, combined with the amount of revision to an already problematic paper, may have been too challenging for some students given the limited time for revisions (~1–2 weeks) at the end of the semester. The paper revision policy may need to be revisited.
The author performed an additional analysis of student projects from the four most popular topics: Hurricanes Andrew and Katrina (each chosen four times), space weather (chosen three times), and the Groundhog Day 2007 tornado outbreak in central Florida (chosen three times). This review encompassed 14 projects (33 total students) and allowed for some “common ground” upon which to judge student performance; the sample included projects from before and after the due-date change. The grades for these groups ranged from 75 to 95, with a mean/standard deviation of 87.0/6.1, close to the overall project average/standard deviation of 85.5/7.4. There was some good student ingenuity in these case studies; two examples from the Hurricane Katrina projects are worth noting. One group suggested that the WIPC be used as a planning tool for future events to identify areas of potential breakdowns between the provider and user communities, whereas another group built their own information flow diagram of the PUR for this case (shown here as Fig. 6). The student PUR diagram is particularly insightful in that it illustrates several sources of weather information for the general public (e.g., NWS, various levels of government, and the media). Despite the dissemination of dire warnings from the NWS, a mandatory order to evacuate New Orleans was not given until less than 24 h before landfall (U.S. House of Representatives 2006). In this case (and illustrated in Fig. 6), multiple information providers could be a source of confusion depending on the message, thus affecting the ability of users to act on the information.
Finally, the student evaluations for the last eight semesters were examined, with particular attention paid to questions about course objectives, organization, and overall satisfaction. Unfortunately, the university changed the course evaluation questions after fall semester 2008, so analogous questions were used to determine student reaction to the course's statement of objectives and its organization. Table 5 shows a listing of the evaluation questions along with the response rates. Generally speaking, the students approve of the way in which the objectives have been presented and organized, with over 70% of the 86 respondents rating the objective/organization in the highest category (questions 1 and 2 in Table 5). The results for the overall rating are also encouraging, with over 90% of student respondents rating the course as above average/excellent and over 80% strongly agreeing that they were satisfied with the course instruction (question 3 in Table 5).
This paper presented a business process model for the weather forecasting enterprise that encompasses both its technical and business operations aspects. The model, adapted from the USAF, is similar to flowchart process models used by Dutton (2002) to describe the transition of the weather forecasting enterprise and those of Morss et al. (2008) and Brooks and O'Hair (2006) that focused on the users of weather information and the outcomes of their usage. The primary purpose of the model in the course is as a central organizing construct, helping students bring together various concepts from their previous coursework and to prepare them for the simulated operations of the Embry-Riddle Applied Meteorology Program's capstone course.
In particular, the importance of the PUR portion of the model cannot be overlooked. It is vitally important for the weather information provider, especially if he or she is in a business–client relationship with a user, to understand user requirements, expectations, and subject-matter grasp of weather and climate. The PUR is also a useful tool for defining roles and relationships within a specific type of operation. For example, as this paper goes to press, a considerable amount of analysis and testing is being done to define policies for production and usage of aviation weather information in the four-dimensional data cube being defined by the Next Generation Air Transportation System program (see, e.g., the summary of the 2011 Friends and Partners of Aviation Weather Summer Meeting, particularly the last session, entitled “Building the SAS, policy and governance challenges”: available at www.ral.ucar.edu/general/Summer_Meeting_2011/).
Assessment of the concept model gave generally good results, as evidenced by above-average final-project grades using the WIPC/PUR model as the analysis template. However, the idea of allowing term paper revisions by changing the deadline to an earlier date appears to have mixed results in that some students may not have fully grasped the WIPC/PUR concept until the end of the semester. One of the ideas being considered is to move some of the PUR material into WX 457, where it may be a better fit with that course's objectives, allowing the students to concentrate more on the technical aspects of the analysis and forecasting process in WX 427.
Although the WIPC/PUR model is a useful teaching tool in this course, its adaptation by educators in other meteorology/atmospheric science programs should be tempered by examining the context within which they intend to use it. This model was helpful in WX 427 because the course is a “waypoint” between the beginning synoptic meteorology course and the weather operations seminar in Embry- Riddle's program. Programs with more than one forecasting course (e.g., practicums, current weather discussions) should evaluate the use of this model only after considering the differences between the way it has been used here and how it may be employed elsewhere.
The usefulness of the model to teach students how to forecast has not been fully evaluated. There are four years of city-pair forecast data that can be analyzed statistically to determine if there is student forecaster learning taking place and to what extent. This type of analysis would be quite useful in linking course objectives with outcomes, in addition to those presented here, and is being considered for future investigation.
I would like to thank several key individuals and groups for their help during the development of the business process model and its application in the undergraduate curriculum at Embry- Riddle. Discussions with Drs. Fred Mosher, Chris Herbster, Thomas Guinn, and Randell Barry in the Applied Meteorology Program were most helpful as I began adapting the model from USAF weather operations to the undergraduate classroom. The students who took the course provided valuable feedback both during as well as after the class for improving the model, the forecasting exercises, homework assignments, and the final project. I learned a great deal from 51 different team presentations covering 35 topics, ranging from the effects of weather on the sinking of the Titanic to discussions of the PUR in the Challenger disaster. The PUR diagram shown in Fig. 6 was originally developed by two of my students, Audrey Kiefer and Marissa Gonzales, and I would like to take this opportunity to acknowledge their ingenuity. I would also like to thank the three anonymous reviewers, whose comments and suggestions for improving this paper were extremely helpful.
The results from the fall semester 2006 were excluded from this study because this was the first time the author taught the course, and it was restructured into its present form in spring semester 2007.