report on a paper entitled : Teaching hydrogeology in the field : the bottleneck in student conceptual model development

 The argument presented in the paper is logic and the motivation for the argument was provided. However, the psychological approach on mental models was not accompanied by factors for such differences among the sample subjects although a reader could deduce such factors based on information presented on table 1 lines 139-143 on page 5. Factor for such prior knowledge could have been elaborated more. For example, does prior knowledge means lessons on groundwater field-school and modelling during undergraduate levels before at masters’ level? This is what it seems to imply.

. Finally, field courses based on a learning-by-doing 85 strategy face difficulties adapting available resources (i.e., site or instruments, between many 86 others) to the students' learning. Therefore, in hydrogeology field courses, it is important to 87 combine in-situ lecture-based explanations (theoretical knowledge) with inquiry-based 88 learning (data gathering, analysis and interpretation). This, along with the acquisition by the 89 students of communication skills with multiple representations, can support the development 90 of higher quality conceptual models. 91

92
The purpose of this study is to assess the importance of prior knowledge in student outcomes 93 from an inquiry-based learning strategy used in a hydrogeology field course, and the 94 effectiveness of the proposed inquiry-based learning in the conceptual model expressions by 95 the students. We hypothesize that the students' prior knowledge would control their 96 performance in the outcome from the inquiry-based learning proposed, and both the prior 97 knowledge and the inquiry-based learning in the conceptual model expression for the 98 experimental site. 99 100 101 https://doi.org/10.5194/hess-2020-206 Preprint. Discussion started: 18 May 2020 c Author(s) 2020. CC BY 4.0 License.

Conceptual versus mental model 102
While a conceptual model is an external and simplified representation based on scientific 103 knowledge of a system and/or its functioning (Norman 1983), a mental model is internal, 104 personal, idiosyncratic and commonly incomplete representation of a system and/or its 105 functioning (Horton, 1915;Greca and Moreira, 2000). The assumption that conceptual 106 models taught to students should be learned by them and used to make a relation with the 107 theory is not necessarily true (Greca and Moreira, 2000). Students bring their mental models 108 both to the theoretical and practical (field and laboratory) lessons, and lecturers often assume 109 that these mental models will evolve into accurate conceptual models (Norman, 1983;Duit 110 and Glynn, 1996). If such evolution is successful, the student then has the correct theoretical 111 basis with which to understand physical phenomena. However, students do not always 112 understand conceptual models, even when the models are presented correctly, because many

Course context and field site 130
The Groundwater Field Course is offered annually as part of the master curriculum in 131 Environmental Engineering at ETH Zürich. In 2019, this four-day (3 days in the field, 1 day 132 in the lab) intensive course included a total of 17 students (7 male, 10 female). While the 133 majority of these students (76%) attended the Groundwater course (theory and modelling) 134 right before the Field Course, the different academic paths followed by each student resulted in a diverse course backgrounds and wide distribution of prior content knowledge between 136 the individual students enrolled (Table 1) Table 1), inquiry-based learning (IBL, see Table 2) and conceptual 198 model expression (CME). Here, we assessed each student's PK based on the number of 199 courses the student had already completed from a list of 14 courses closely related to 200 hydrogeology (e.g., surface hydrology) and visual representation (e.g., GIS). Each student 201 thus was assigned a PK rank of either low (≤ 6 courses) or high (≥ 7 courses). The remaining 202 two variables were quantified based on each student's scores. Each student's success at 203 inquiry-based learning (IBL) was quantitatively evaluated from a written report [0-6], and 204 https://doi.org/10.5194/hess-2020-206 Preprint. Discussion started: 18 May 2020 c Author(s) 2020. CC BY 4.0 License. success at conceptual model expression (CME) was quantitatively evaluated from the 205 conceptual model representation [0-6]. With these variables, we then sought to answer our 206 questions with two types of statistical tests: descriptive and inferential statistics for one 207 sample and a test of statistical relationship between two samples. 208 209

Results and discussion 214
To determine how prior knowledge may affect a student's success in our Field Course, we 215

240
To assess the control that IBL exerts on CME, we performed a regression analysis and the 241 Mann-Whitney U test within the two PK groups (high and low) ( Table 5). Although no 242 correlation was observed between IBL and CME in both PK groups, a significant relationship 243 between IBL and CME was found in both PK groups. The low R 2 values (0.066 and 0.093 for 244 high and low PK groups, respectively) indicate weak or no correlation between IBL and 245 CME scores, at least for a linear model. However, the low p-values (< 0.05) obtained still 246 reflect a real relationship between the predictor (IBL) and the response variable (CME). In 247 other words, there was a positive relationship between the scores in the conceptual model 248 expression and the performance in the inquiry-based learning, with higher average scores in 249 the high prior knowledge group than in the low. 250

251
In summary, student scores were overall lower for conceptual model expression than for 252 inquiry-based learning. However, high variability in CME scores was observed between 253 students in the high PK group. Some students with a high rank in prior knowledge still scored 254 poorly at conceptual model expression, indicating that a larger number of previously attended 255 courses does not guarantee a better performance at model conceptualization (though in 256 general higher CME scores were associated with high PK). Similarly, high variability in IBL 257 scores within the low PK group indicates that prior knowledge does not correlate to strong 258 performance in the field (performing experiments, data collection and analysis).   Course students were attending a groundwater theoretical and modelling course previously).
As demonstrated in this study, inquiry-based learning does not directly contribute to the 295 success of conceptual model expression. Thus, specific lessons in the classroom (prior to 296 going to the field) to introduce methodologies for conceptual model expression should be 297 integrated into courses based in active learning. In addition to the development of 298 computational skills (i.e., spatial reasoning abilities), the inclusion of physical models for 299 classroom teaching (prior to going to the field) would aid student elaboration of conceptual 300 models and provide a connection between theoretical knowledge and reality (Rodhe, 2012). 301 This pedagogical combination would aid students as they learn how individual field 302 experiments and the information provided by them can contribute to conceptual models and 303 how we express them.