Portal de Periódicos da CAPES

Engaging secondary school students in climate data rescue through service‐learning partnerships

2020; Wiley; Volume: 76; Issue: 4 Linguagem: Inglês

10.1002/wea.3841

1477-8696

Carla Mateus, Aaron P. Potito, Mary Curley,

Species Distribution and Climate Change

Many meteorological records remain as paper data sources, which constitutes an obstacle to climate research. Student–scientist collaborations between secondary schools, universities, national meteorological services, and archives have the potential to be applied elsewhere and contribute to a fast, accurate and low-cost climate data rescue process. Service-learning partnerships in climate data rescue projects are crucial to communicate climate science and to enhance climate data literacy outside the school curriculum. Secondary school students–scientists partnerships are an important new avenue for climate data rescue projects. Long-term instrumental series are necessary for a greater understanding of past climate variability and trends and to better assess the frequency, magnitude and duration of past extreme weather events (Klein Tank et al., 2009). Historical meteorological observations are needed for the development of climate products such as gridded datasets and reanalysis (Stickler et al., 2014). Moreover, long-term climate series are essential to validate palaeoclimate reconstructions from proxies or documentary sources; to corroborate climate modelling; to support climate change detection and attribution studies; and to assist climate action, mitigation and adaptation policies (World Meteorological Organization, 2016a; 2018). Numerous meteorological records dating prior to the mid-twentieth century remain as unique images or as vulnerable paper data source formats (World Meteorological Organization, 2016b; Brönnimann et al., 2018) despite many climate data rescue efforts such as the International Atmospheric Circulation Reconstructions Over the Earth (ACRE; Allan et al., 2016) and The International Data Rescue (I-DARE) Portal (https://www.idare-portal.org/). Climate data rescue is necessary to preserve the heritage of meteorological observations from being lost and to extend the temporal and spatial availability of instrumental data (Allan et al., 2016; World Meteorological Organization, 2016b). This lack of digital and open-access meteorological observations constitutes an obstacle to climate research. Thus, open-access digital data for researchers, national meteorological services, stakeholders and policymakers is necessary to fill key gaps in climate research. Old and fragile meteorological manuscript registers may contain diverse legibility issues such as blurred or poor handwriting, faint ink, corrections made over previous registered readings, omission of points as decimal separations, reversed observations or values written in an incorrect climate element column. Distinct methodologies for the digitisation of climate manuscript data, including software for speech recognition, optical character recognition and manual keying (of keystrokes on a keyboard), have been tested by Brönnimann et al. (2006). Legibility of data sources may compromise the accuracy of data digitisation methods such as the optimal character recognition, and thus, manual keying is considered the best feasible method for climate data digitisation (Brönnimann et al., 2006; World Meteorological Organization, 2016b; Ashcroft et al., 2018). Double (or more) data keying is recommended to minimise transcribing errors during data rescue (World Meteorological Organization, 2016b). However, climate data rescue can be a time-consuming and expensive task, requiring the hiring of extra staff for data keying, which can be an impediment for the digitisation of climate data. Recent advances in climate data rescue incorporate crowdsourcing by volunteer citizen scientists for manual data keying through web-based tools as a fast and efficient methodology to digitise large amounts of observations such as Old Weather, 1 Rainfall Rescue, 2 Data Rescue: Archives & Weather 3 or Weather Rescue (Hawkins et al., 2019). The enrolment of university students in climate data rescue as part of an assignment was inspired by crowdsourcing efforts and is reported as being as accurate as those conducted by professionals (Ryan et al., 2018). In addition, small groups of university students have already cooperated with national meteorological services and researchers as part of paid contracts (Ashcroft et al., 2018). The engagement of secondary school students in climate data rescue activities through non-mandatory service-learning partnerships between secondary schools, universities and meteorological services has, to date, been unexplored as a fast and efficient science outreach methodology to accelerate climate data rescue. Service learning is an experiential teaching and learning pedagogy in which students apply the knowledge and skills learned in the classroom through the engagement in systematised activities with clear goals to fulfil community needs (Sigmon, 1979). Data collection as a part of service-learning partnerships has been applied as a tool to integrate climate change and sustainability concepts in the classroom (Coleman et al., 2017). Furthermore, hands-on instruction and practice allowed a network of secondary school students to be enrolled in data collection under student–teacher–scientist partnerships such as the Global Learning and Observations to Benefit the Environment (GLOBE) programme (Creilson et al., 2008). These student data collection partnerships have been described as being of sufficient quality and accuracy for scientific research and have therefore been promoted for use by the research community (Congalton and Becker, 1997). Moreover, scientific data collected by secondary school students as part of student–scientist partnerships have been characterised as reliable and consistent with data taken by professionals such as by Hydrometeorological Institutes (Mims III, 1999). In this research, two methodologies have been investigated: hosting and training students at university and training students at a secondary school under a student–teacher–scientist partnership. Opportunities in climate data rescue were circulated to the directors of the secondary schools in Co. Galway, Ireland. A total of 127 students from seven schools were hosted at the National University of Ireland Galway (NUI Galway) for periods of 1 day to 1 week. The students were accommodated in small cohorts, typically of 15–20, from 10h to 16h with breaks for outreach activities. One of the schools engaged 17 students in the service-learning partnership twice as the students expressed an interest in returning for further climate data rescue activities. The student–teacher–scientist partnership also enrolled 18 students from one school in Co. Dublin as part of the Green School Module, which is the Irish application of the international sustainable education programme Eco-Schools. 4 The 18 students completed data keying over 3 weeks with classes of 40 min twice a week supervised by the trained teacher. Contact with the teacher was established by email when required. The 145 students, aged 15 and 16 years old, were registered in the 10th grade, which corresponds to the Transition Year Programme – an optional non-academic year between the examinations in the Junior Cycle (grades 7–9) and the Senior Cycle (grades 11–12), which is unique to the secondary schools in the Republic of Ireland. The Transition Year Programme aims to clarify career pathways, to enhance transferable and practical skills needed in a working environment and to create opportunities for students to apply the skills and knowledge learned in the classroom. Training and induction for students and volunteers included a review of the research project aims and outcomes, the importance of climate data rescue and the purpose of double keying for data digitisation accuracy. Specifically, the significance of long-term instrumental series to understand past climate variability, trends and the assessment of extreme weather events was also explained. In addition, the value of the rescued data to support climate change detection and attribution studies, to assist climate modelling and to generated climate products was discussed. The rich heritage of early instrumental meteorological observations in Ireland was reviewed, and an overview of station metadata was presented, including station drawings and photographs (Figure 1). Station metadata, such as the diverse types of thermometer screens in use prior the introduction of the Stevenson screen, were highlighted. Images of data sources such as handwritten registers, newspapers and proceedings of societies were examined for climate data and metadata. All secondary school students were given a welcome talk on the diverse research clusters in geography at NUI Galway and given information on possible career pathways. Students hosted at NUI Galway were also provided with a campus tour, which included a visit to the library and natural history museum facilities. Additional outreach activities included talks and inquiry-based demonstrations by staff members at the laboratories in geography on palaeoclimate research. Double keying is a necessary procedure to minimise climate data digitisation errors (World Meteorological Organization, 2016b). Under a research project for the assessment of past extreme air temperature events in Ireland, the first author completed one data keying of daily maximum and minimum air temperature observations from 12 long-term and 21 short-term series dating back to the early and mid-nineteenth century (Figure 2). The practical training on climate data rescue skills consisted of a demonstration on how to perform manual data keying from diverse data sources into MS Excel templates and guidance on how to avoid digitisation errors. Each participant was requested to make a visual cross check between the keyed data and the image of the data source to help identify any digitisation errors and to confirm the correct number of keyed calendar days per month. Moreover, each participant was required to compare the monthly average and sum generated in the MS Excel template with the monthly average and sum in the data source. The participants were furnished with station folders for data keying. Each station folder comprised images of the original data source containing daily maximum and minimum air temperature observations, as well as other climate elements, for a year for each station. In addition, each folder included a MS Excel template for data keying and instructions for the identification of the thermometer observation columns in the image and how to complete data input into the MS Excel template. Students (and volunteers) were required to identify daily maximum and minimum air temperature values per calendar day in the image of the original data source and to input them as they are written in the data source following the standard data rescue practice advised by the World Meteorological Organization (2016b). The MS Excel templates (Figure 3) were basic and similar to the original data source (Figure 4) in order to minimise data-keying errors and to allow fast keying. Benchmarking with previously explored methodologies was established to compare groups and to evaluate the data-keying accuracy of secondary school students. A total of 193 first-year BA Joint Honours (Geography) and BSc Applied Social Science students fulfilled data rescue in the module Geography in Practice as part of an assignment on statistical data analysis similar to that presented by Ryan et al. (2018). A total of 12 NUI Galway students participated through the volunteering programme ALIVE (A Learning Initiative and the Volunteering Experience). In addition, seven members of the Irish Meteorological Society were included in the benchmarking. MS Excel macros were created to compare the consistency of the primary data keying undertaken by the first author of this research and the secondary data keying completed by students and volunteers. In case of a difference, the second author of this project was provided with an image of the data source and the date on which there was difference in order to confirm the correct value. Anonymous five Likert scale (5 = strong yes, 4 = yes, 3 = neutral, 2 = no and 1 = strong no) questionnaire-based surveys were distributed to secondary school students, university students and volunteers at the end of the data rescue activities to investigate self-reported knowledge and perceptions on the teaching and learning experience of the student–scientist partnership. The secondary school students engaging for a second time in climate data rescue activities repeated the survey at the termination of the service-learning partnership in order to determine any changes in the students' perceptions. Qualitative feedback provided in the optional survey comments was also examined through content analysis. A total of 357 participants rescued 775 582 daily air temperature observations (Table 1). The 127 secondary school students hosted at NUI Galway rescued the greatest proportion of data, with a total of 418 116 values keyed in 1005 templates. However, the 19 individual volunteers were responsible for a greater average contribution, with 11 811 rescued observations per volunteer. The cross checking of keyed data allowed for the identification of errors (0.036%, or about 1 in 3000) in the first data keying. Of the second keying groups, volunteers showed the lowest rate of keying errors (median of 0.3% errors per MS Excel template; Figure 5). Approximately a third of the volunteers are members of the Irish Meteorological Society who have a professional background or interest in climatology or meteorology, which may explain the extra care and low rate of errors. The secondary school students hosted and supervised at NUI Galway had a slightly lower error rate (median of 1.6% errors per MS Excel template) in comparison to the students supervised by the trained teacher at the school through the Green School Module (median of 1.8% errors per MS Excel template). The greatest data rescue errors were found in the templates filled by the undergraduate students (median of 3.3% per MS Excel template), which may be related to the mandatory engagement in climate data rescue as part of a graded assignment. Student t-tests on the percentage of errors per MS Excel template indicate that the volunteers as a group have significantly lower errors than the other groups (P < 0.001). Furthermore, the ALIVE volunteering students committed significantly fewer errors in comparison to the undergraduate students registered in the data rescue assignment (P = 0.008), which may be related to the non-mandatory engagement in climate data rescue for the volunteers. Although the median of the percentage of data rescue errors per MS Excel template is lower for the secondary school students than for the undergraduate student assignments (Figure 5), the difference between these groups is not significant (P = 0.216). Regarding the perceptions on teaching and learning, the secondary school students recognised the importance of the digitised data for service-learning partners: 87% recognised that the data are important for academic research, and 85% consider the data crucial for Met Éireann (The Irish Meteorological Service) (Figure 6). Of the students, 88% enrolled for the first time in climate data rescue remarked it as a positive experience. The students engaged for the second time in the student–scientist service-learning partnership showed greater positive perceptions on the importance of climate data rescue (Figure 6). Overall, 50% of students reported an interest for future climate data rescue activities, and 42% of students would like to pursue a career in geography or meteorology. Interestingly, 71% (n = 90) of the students hosted at NUI Galway were girls, of which 17 were engaged twice in climate data rescue, which shows that young female students are particularly interested in science, technology, engineering and mathematics (STEM) fields. The content analysis of the optional survey comments (Table 2) demonstrates that several secondary school students acknowledged the acquisition of new competencies in addition to the gained climate data rescue skills. For example, one comment read as follows: 'Overall the exercise was very good and even if I do not pursue a career in Meteorology, I can now use Excel, which is a very useful skill to have'. Student–scientist research partnerships as part of non-mandatory service learning on climate data rescue are crucial to communicate climate science, to engage secondary school students with climate data and thus to enhance climate data literacy outside the school curriculum. Climate data collected by secondary school students engaged as research collaborators are of benefit to scientific research. Service-learning partnerships between secondary schools, universities, national meteorological services and archives have the potential to be applied elsewhere and contribute to a fast, accurate and efficient climate data rescue process. Student–scientist partnerships can be practiced in countries without similar programmes as the Transition Year Programme. Specifically, research partnerships have the potential to be developed during science week outreach events. Secondary school students could be hosted and trained by climate scientists at universities, schools, national meteorological services or archives to acquire climate data rescue skills and to gain insights into climate science while exploring career pathways. Engaging secondary school students in climate data rescue under student–scientist partnerships through web tools could allow students from diverse geographical areas to be included in climate science outreach projects while contributing to a fast climate data rescue process and could increase the volume of data rescued. Climate data rescue activities could be extended to all Transition Year (and other secondary school) students in Ireland as part of a science week project through a web-based approach with online video instructions, and a discussion forum. The authors recommend meaningful student–scientist research partnerships as part of organised climate data rescue projects, with clear climate science learning goals and outreach activities to enhance the students' interest in climate science and to highlight careers in STEM fields. The accuracy rate of secondary school students (95.2%) is comparable to citizen science data rescue projects such as the Weather Rescue, which highlights an agreement of over 95% between a minimum of three data entries (Hawkins et al., 2019). Due to the benefits for students and researchers, as well as the high accuracy of rescued data, the authors advise that service-learning partnerships between secondary school students and scientists should be considered an important new avenue for climate data rescue. The daily maximum and minimum air temperature dataset rescued in this work is available as open access in Mateus et al. (2020). This research has been funded by the Dr Tony Ryan Research Scholarship from NUI Galway. The authors thank climate data rescue by: secondary school students (Transition Year) from Calasanctius College Oranmore, Claregalway College, Dominican College Taylor's Hill, Galway Community College, Our Lady's College, Presentation College Athenry, Salerno Secondary School, and Holy Child Community School; NUI Galway first-year BA Joint Honours (Geography) and BSc Applied Social Science undergraduates; members of the Irish Meteorological Society; Met Éireann volunteers; and NUI Galway students through the programme ALIVE in the academic year 2017/2018. The authors are grateful to the Editor and reviewers for their insightful suggestions.