DOI: 10.31038/GEMS.2023567

Abstract

The paper introduces a novel, Socratic approach to creating research studies, combining the emerging science of Mind Genomics with AI embedded in the Mind Coach system. The templated system instructs the researcher to describe a topic. AI (Idea Coach) synthesizes sets of 15 questions based upon the topic description, and then requires the respondent to select four of the questions. Afterwards, the system instructs the respondent to create four answers to each question, once again providing AI help through the suggestion of 15 answers. This first Idea Coach steps can be repeated as many times as the researcher desires, with each step returning 15 questions, many different, along with a later and automatic synthesis of patterns in the 15 questions through seven standard AI queries. The experiment with actual people presents each study participant with a set of 24 vignettes, combinations of the aforementioned answers. Each respondent evaluates a unique set of 24 different vignettes created by an underlying permuted experimental design. The ratings are analyzed by dummy variable regression analysis to show how each element or answer ‘drives’ the rating. The respondents are further clusters into like-minded groups (mind-sets), based upon their coefficients. The templated approach creates a rapid (2-4 hour) process to go from a state of almost zero knowledge to understanding a topic through AI, geography in this particular instance, to produce the test stimuli, and through human reactions to these test stimuli determine and describe the mindsets respondents hold towards geography that can be organized into courses of specific interest to each mind-set.

Keywords

Mind genomics, AI, Socratic method, Question creation

Introduction

In the “project of science,” research studies are assumed to be efforts which contribute to a picture of “how the world works.” The assumption is that the researcher can identify what might be the next experiment to perform. The experiments often end up as simple reports, supported by statistics, and introduced by detailed literature reviews. Those who publish their investigations are often described as “filling gaps in our knowledge.” Indeed, the much of the edifice of science rests on the practice of what is called the “hypothetico-deductive” system, the system which requires that the researcher propose a hypothesis and do the experiment to either support the hypothesis or falsify it. It is by the accretion of such studies that the edifice of science is created, the picture of the world [1], a picture created by a disciplined approach.

At the other side of the project of science is grounded theory [2,3]. Here the researcher does a study or reports a set of observations. It is from those observations that hypotheses emerge. Once again, however, the effort assumes at the start that the researcher does the experiment, and thus implicitly assumes that the researcher is beginning with a knowledgeable conjecture.

What then happens in those increasingly frequent cases where the issues are new, or at least new combinations of old issues, and where there has not been sufficient time to create a literature, or even to develop grounded theory and hypotheses? Can a method be developed which allows the exploration of issues in a manner which is quick, simple, yet profound in the depth of information and insight that it can promote, and even create? This paper presents such an approach with a worked example, and a timetable of events. The topic is creating an understanding one can create a course in geography, starting with little knowledge, and obtain early-stage feedback. The objective is to test out a general system, Mind Genomics, which has evolved over the past 30+ years, keep pace with developments in statistical computation, and the trend to DIY (do-it-yourself). The impetus for the focus on geography was reading about the launch of a new journal on geography, and the author’s own experience with geography and earth science during the formative high school years. The literature focusing on the teaching of geography provided some of the necessary background [4-6]. Further impetus from the study came from work on horticulture [7] and on travel to Albania and the focus on the pleasant weather of Albania [8].

The approach presented here, Mind Genomics, comes from a combination of three disciplines, and has evolved since the 1990”s [9,10]. The disciplines are:

1. Mathematical psychology and psychophysics; The study of how we subjectively “measure external stimuli and situations in our mind,” to create an algebra of the mind. For the current topic of geography, mathematical psychology and psychophysics will help us create the structure of how we think about topics.
2. Statistics, specifically experimental design. This is the study of how we can combine different variables to represent alternative “realities,” these realities equivalent to different descriptions of how the world works. The normal, everyday experience of the world comes in packets of stimuli, not in single ideas. Rather than surveying the person, giving that person single questions, we create combinations of those questions, and give the person these combinations. The person then rates the combinations. It will be from the deconstruction of the responses to these mixtures of ideas into the contribution of the individual “elements” that we will learn about the mind of the respondent.
3. Consumer research. Consumer research focuses on the world of the everyday, the normal. Rather than creating an artificial situation to elucidate an issue, putting the subject in the situation and then observing behavior, the consumer researcher looks for the world of what people typically experience.

Mind Genomics emerged as a science of the everyday, spurred on by the aforementioned three sciences to measure how people think about ordinary experience. The objective of Mind Genomics is to measure the value of an idea, following the spirit of traditional psychophysics, which focuses on measuring sensory experience. In other words, can we measure the “intensity” of an idea?

Mind Genomics moves away from the conventional approach of surveys, which instruct survey-takers (henceforth called respondents) to rate the importance of different factors, using a scale. The typical use of surveys is to focus a person’s mind on one aspect at a time, one experience, or one’s attitudes regarding a topic. Typically, the survey instructs the survey taker to think of large aspects of the experience and rate the aspect as a totality. Sometimes the desired focus is finer. For example, by dividing the topic into small, “bite-size” ideas of single focus (e.g., aspects of health in a polluted area), a survey can have the survey taker quantify many of the simple aspects of the experience.

The problem of surveys, and the stepping off point of Mind Genomics, is that the experience is not a set of individual aspects but rather combinations of aspects. People do not naturally respond to single ideas, although when instructed they can intellectualize their experience, and come up with a best guess. The typical survey situation is sterile, lacking the richness of experience because the phrases in the survey are “general.” If one were to summarize the thinking involved in survey development, it would be “top down,” wherein the researcher wants the survey taker to abstract, rather than deal with the granular richness.

Mind Genomics works from the bottom up, from the world of the granular, not from the world of the general. A Mind Genomics study first creates a set of phrases describing the granular aspects of the topic. It is impossible, of course, to capture all of the granular aspects of an experience or topic, so the operationalized approach is to focus on four aspects of a topic, each aspect “brought to life” with four granular and different statements. The aspects can be considered “questions, the statements can be considered “answers” or “elements. In other words, the Mind Genomics process thus assembles four questions, each with four answers, hopefully the questions providing a story.

The execution of the Mind Genomics is straightforward. The researcher creates the questions, the researcher creates the answers, types both of these into a template (www.BimiLeap.com), along with a rating question. The BimiLeap program then creates 24 different combinations of answers for each respondent. These combinations are called “vignettes,” and comprise 2-4 elements, at most one element or answer for each question, but often no answer from one or two questions.

Each respondent ends up evaluating virtually a totally different set of 24 vignettes. Each set of vignettes is set up to be structurally identical to every other set of 24 vignettes, at least on a mathematical basis. The differences among the various sets of 24 vignettes are that the elements are permuted. For example, considering respondent (survey taker) #1, A1, A2, A3 and A4 correspond to four different elements. For the second respondent #2, permutation goes into effect. The element previously A1 becomes element A3, the element previously A2 becomes A4, the element previously A3 becomes A1, and the element previously A4 becomes A2. Each set of vignettes is constructed to ensure that the 16 elements are statistically independent of each.

The subsequent processes explicated below, comprise presenting these test vignettes to respondents, obtaining ratings, transforming the ratings to create new “binary variables”, and finally using both regression analysis and to reveal how people perceive different aspects of a topic, and how people end up falling into different groups, so-called “mind-sets”, operationally defined as groups of people who think of the topic in quite different ways.

Difficulties Encountered in the Creation of Ideas, and the Help Provided by Artificial Intelligence

During the evolution of the Mind Genomics platform from its beginnings in the 1990’s under the name IdeaMap, one behavior has continued to repeat study after study, especially among the beginners using Mind Genomics. Simply stated, the difficulty is coming up with the four questions, and to a lesser degree coming up with the four answers for each question. Even accomplished researchers unfamiliar with the Mind Genomics approach report difficulty during the first time that they try this exercise. Quite often the difficulties encountered in the sheer effort of “thinking through the problem” suffice to so strongly discourage that the researcher abandons the task. On the other hand, one the researcher pushes through and creates several sets of questions, the discouragement vanishes, often leading to curiosity and excitement.

The story is a bit different for the second part of the effort, providing answers to the questions. Creating the questions was the hard part. Providing four answers is the easy part, perhaps because we are taught to answer questions, not to pose them. The only thing one has to learn when creating answers is to write the answers in such a way that they can be scanned, almost like a menu.

The remainder of this paper deals with the creation of the relevant set of questions through the Idea Coach, and the explication of the output provided to the researcher, both through inspection of the results, and through AI-enhanced summarizations.

Method

The actual process of creating the test elements begins with the request to provide four questions (Figure 1, Panel A). It is as this point that the researcher needs the help of the provided by the AI in the Idea Coach. Figure 1, Panel B shows the request to Idea Coach, inserted by the researcher.

Figure 1: Panel A presents the request to provide four questions. Panel B presents the paragraph written by the researcher to direct the AI powering Idea Coach.

Table 1 presents the first set of 15 questions generated by the Idea Coach. The Idea Coach is set up to allow the researcher to select from no questions to all four questions from this first “run” of Idea Coach. The Idea Coach can run many times as the researcher searches for the best four questions. For each “run” the Idea Coach returns with some new questions, and some old questions. Each run of the Idea Coach is separate, unconnected with the previous run. The researcher is free to edit the question before accepting it, and even to write some or all herself or himself.

Table 1: Questions emerging from the first run of Idea Coach. The Idea Coach is presented with a topic, and instructed to return 15 questions.

An additional feature of Idea Coach is the use of AI to summarize the questions for each set of 15 questions. Table 2 shows a partial table of this additional summarization. Each time the Idea Coach is run to generate questions, the 15 questions generated are summarized anew. As a happy consequence, the researcher who runs the Idea Coach 5-10 times to generate questions will also receive a detailed summarization of the topics from the AI embedded in Idea Coach.

Table 2: Some AI-based summarizations of the first set of 15 questions. In the interest of space only the first two or three items from each summarization are shown.

Once the researcher has generated the four questions, each question appears, one at a time, in a format similar to Figure 1, Panel B. The researcher is instructed to provide four answers to the question. The Idea Coach, viz., the AI, provides sets of 15 answers to the researcher who once again must select four answers for each question Figure 2 shows the process. Panel A shows the text of the first question, and the four spaces for the answer. Panel B shows the request to Idea Coach for the four answers. There is no need for a “box” for the researcher to write in the text of the question since that information is already in the system from the creation of the questions. Panel C shows some of the output from Idea Coach. Panel D shows the four answers to Question #1, after they have been received from Idea Coach, selected to be put into the program, and then edited if necessary by the researcher to make sure the answers are in the proper format, viz., a meaningful phrase or better a meaningful and simple sentence.

Figure 2: The process followed by the Mind Genomics program to generate and select answers to a question, using Idea Coach.

The Idea Coach can be used many times, allowing the underlying AI to produce different sets of 15 questions, each time responding to a single “squib” or problem statement, as well as different sets of 15 answers to the same question. Across the repeated set of questions or answers, some questions or answers will repeat, but others will be new. Each set of 15 returned questions or returned answers to a question is further summarized by AI, according to a variety of queries to the Idea Coach. It simply requires that the user formulate the questions and answers. For this example, and in a period of less than 15 minutes, it was possible to set up the program, type in the description in Idea Coach, iterate through three sets of 15 questions each, select the four questions during the course of the four iterations, and then once again use Idea Coach to provide the four answers to each question, with three iterations chosen for each question. The result is that within 20-30 minutes it was possible to create a set of 15 such inquiries, with each inquiry comprising 15 “returns” from the AI embedded in Idea Coach. Afterwards, the Mind Genomics program emailed the researcher with the “Idea Book” of 15 pages, one page for each set of 15 questions or answers produced by Idea Coach, as well as additional AI-based summarizations of the 15 questions and answers.

The actual study proceeds with the creation of the classification scale, and then the rating scale and the anchored scale points. Table 3 presents the specifics of these scales, along with the orientation that the respondents read. It is important to keep in mind that for Mind Genomics the vast amount of information comes from the response to the particular elements. The orientation simply tells the respondent what to do. Generally, the orientation is “skimpy,” requiring the respondent to use the information in the test stimuli when they make their rating.

Table 3: The self-profiling classification scale, the respondent orientation to the Mind Genomics evlauations, and the rating used to evaluate the test vignettes.

The Respondent Experience, and the Construction Strategy Underlying the Vignettes

After the research finishes the study set-up, following the template, choosing questions/answers, creating the self-profiling questions, the orientation to the respondent and the rating scale, the research launches the study. The researcher can choose the source of respondents, either from panel providers who specialize in “on-line” research, or from other sources, such as colleagues or students.

Figure 3, Panel A shows the first screen emerging after the welcome to the study. The panel shows the self-profiling questions as a series of pull-down menus. The appearance on the screen of this self-profiling questionnaire is deliberately made to be as “clean” as possible, in order not to intimidate the respondent.

Figure 3: Panel A shows the drop-down menu of the self-profiling classification. Panel B shows one screen shot of a test vignette shown to the respondent, along with the question and the 5-point rating scale.

The next set of 24 screens is the vignettes, shown as an example in Figure 3, Panel B. The screen appears to be a randomized combination of four answers from Table 1. The combination is called a “vignette.” The respondent is instructed to read the vignette, consider all the elements of the vignette to be part of the same idea, and then to rate the vignette on the 5-point scale, following the instructions provided by the rating scale. Note that the vignette has no questions, just answers (elements). Nor does the vignette have any logical or even forced connection between the four elements.

To virtually all participants in these Mind Genomic studies the test stimulus shown in Panel B seems to be random. When faced with a set of 24 such test stimuli, one after another, with no seeming theme, many respondents feel that they have just gone through a set of random combinations. Indeed, the word “random” is often used to describe these vignettes. It is simply impossible to detect a pattern.

The power of the underlying experimental design comes from the way it is constructed.

1. All elements appear equally often.
2. The combinations for each respondent are set up to be a complete experimental design. That is, the researcher can use the data from one respondent to create an equation showing the contribution of each of the q6 elements or answers.
3. Each respondent evaluates a totally different set of 24 combinations. This is known as a permuted experimental design [11]. The benefit of the permuted design is that the researcher gets to cover a great deal of the possible design space. The practical outcome is that the researcher need not know anything about the topic when the research begins. The researcher ends up “exploring” the topic from many angles. The analogy here is the MRI, used to take multiple snapshots of an item from different perspectives, and then combine these snapshots to create a three-dimensional image.
4. Each of the 16 answers or elements appears precisely five times in the 24 combinations and is absent from 19 of the vignettes.
5. Each vignette comprises a minimum of two elements and a maximum of four elements.
6. One question can contribute at most one answer or element to a vignette. This prophylactic ensures that the vignette does not contain two mutually contradicting statements, different alternatives or answers from the same question,
7. Panel B of Figure 3 shows the sparse design of the vignette. The questions do not appear. Only the combinations of answers appears, along with the rating question and the rating scale. Often those who commission the research want to have more “fleshed out” vignettes, with connectives and with sentences. The objective of Mind Genomics is to understand what the specific elements are, or really specific ideas, which drive the response. It is easier for the respondent to be given series of vignettes similar to those shown in Panel B. The respondent ends up grazing, looking at the ideas, and then assigning a rating
8. The foregoing strategy ends up with different combinations of elements evaluated by the respondents, with the analysis valid at the individual level as well as the group level. It will be the power of analysis at the group level which will enable the researcher to discover important patterns, even in those all-to-common situations where the researcher is a beginner, even a school-age student.
9. The final benefit to be surfaced here is the ability of the experimental design to frustrate attempts to “game” the system. Post-study comments by many respondents, often unsolicited, is that they could not game the system, that the system seemed entirely random, that the same elements kept appearing, and that after one or two vignettes were evaluated the respondent claimed to have lost interest and simply guessed the answers. Although one might be perturbed at this admission, especially professionals who want the project to be totally “rational,” the loss of interest means that the respondent reacts at almost a “gut level” to the vignettes, in the same way that the respondent reactions at a relaxed, non-involved way, to the outside world. It is this loss of focus on the “right answer” which allows Mind Genomics to identify how people really think about the topic, rather than spend their time trying to outsmart the researcher.

Relating the Presence/Absence of the Elements to the Rating

The aim of Mind Genomics is to measure the impact of the elements on a scale, or in this study actually on two scales. The five-point scale shows two rating points recording that the respondent is interested (rating 5 and rating 4, respectively), and two rating points recording that others are interested (rating 5 and rating 2, respectively). Furthermore, the five-point scale shows two rating points recording that the respondent is not interested (rating 2 and rating 1), and two rating points recording that others are not interested (rating 4 and rating 1). Finally, rating 3 shows that respondent simply cannot answer those questions.

To link the ratings to the elements requires that we create new scales, called “binary transformed scales.” These are not metric scales of “amount” yes/no scales. The transformations are shown below. They allow the researcher to understand how each of the elements drives interest or disinterest. The transformations are followed by the addition of a vanishingly small random number, that number added to each transformed variable. The addition of the vanishing small random number is a prophylactic step to ensure that every one of the newly created five binary variables shows some marginal or higher degree of variability.

After the ratings are transformed into the five new binary variables, one can run an OLS (ordinary least squares) regression, relating the presence/absence of the elements to the binary transformed rating. The regression will work for each of the five binary variables because the vignettes at the level of each respondent were created to conform to an individual-level experimental design. Thus, the 16 elements (A1-D4) are statistically independent of each other and set up for OLS regression (Craven & Islam, 2011).

The particular “flavor” of the OLS regression is called dummy variable modeling (Hardy, 1993). The term “dummy variable” comes from the coding of the variable. Rather than “amount of,” which is meaningless in this analysis, the coding refers to “presence” (coded as 1), or absence (coded as 0).

The equation is expressed as: Binary Variable = k₁A1 + k₂A2 … k₁₆D4

From various studies with the foregoing equation, one lacking the additive constant, the general operating procedure has been to look at elements of coefficient lower than 2 (viz., 1, 0, negative) as “irrelevant,” to look at coefficients of 21 or higher as “important”, and the coefficients between 2 and 20 as interesting, perhaps as part of a story. These operational rules come from both statistical regression modeling and observations of how these elements seem to perform in the daily world.

Table 4 shows the coefficients for the total panel, with the five binary transformed variables, as well as with response time as the sixth dependent variable. Strong performing elements, those with coefficients of 21 or higher, have been shaded. Table 4 suggests a fair number of elements end up interesting the respondents, all four from Question B (Explain how understanding geography helps explore current events linked to geography and the physical world), and three of the four from Question C (Go into detail about learning about scientific methodologies and tools used in geographical research). In contrast, no elements from Question A or Question D drive interest.

Table 4: Coefficients of the equation relating the presence/absence of the 16 elements to the newly created binary dependent variables.

The same search for strong performing elements may be done for the other binary transformed variables. Respondents in this study did not know how to respond when the question involved other people, and when the question involved disliking the element. We could stop there, were we to be focusing on the research as a way to help us decide what to feature in this publication on geography. Or we could move on to hypothesize that it may not be productive to ask people about what other people think, or even to ask people what they themselves dislike.

Table 5 continues the analysis, this time looking at the way people describe their interest in geography. The respondents were instructed to check one of four types of interest in geography. Only two of the four classifications were chosen by 10 or more respondents; It’s ok… i know some… but really only the stuff i read in the press, and I know to understand some of the fine points. Table 5 shows that there are some elements which appeal far more strongly to one group than the other, but the strong differences in patterns of preference among the elements fail to materialize.

Table 5: Performance of the elements by two self-defined groups of respondents, the criterion being to describe the nature of their interest in geography.

Beyond WHO to Mind-sets and to Measures of Overall Performance Enhanced by Mind-sets

If these data were to be simply type of messages, but without the content of the messages, the researcher might be happy to discover that some messages perform strongly among the total panel, but that different messages perform strongly by how the respondent describes his or her interest. Tables 4 and 5 provide meaningful data, as least at the level of specifics. The data point out which messages or elements are important. The meaning of the messages tell us even more, namely the content of the messages which are important to the total panel versus the groups. What is missing, however, is a powerful unifying pattern, one which need not be synthesized, but rather is compelling. In the words of Harvard’s Professor of Psychophysics, SS Stevens, the patterns should emerge with “ocular trauma,” namely they should hit you between the eyes with their clarity!” [12].

Unifying patterns emerge when the Mind Genomics data are analyzed with an eye towards discovering groups of people who think about the topic in a similar way. Keep in mind that the Mind Genomics ‘program’, viz., the Mind Genomics effort, focuses on the world of the granular, the world where people feel comfortable because it is the world of the everyday. With that in mind, the search for these mind-sets ends up producing far clearer patterns, and in turn, ends up generating a stronger data set, as will be shown.

Mind-sets are defined as groups of individuals who show similar patterns of thinking about a granular topic, these similar patterns operationally defined as similar patterns of coefficients. Each respondent generates a set of 16 coefficients, based upon the equation relating the presence/absence of the elements to the binary dependent variable. For this study, the binary dependent variable was selected as ‘For Me.’ Viz., ratings 5 and 4 transformed to 100, ratings 1, 2, and 3 transformed to 0.

The actual statistics fall into the category of clustering [13], the specific form being k-means clustering. Through k-means clustering, the statistical program operationally defines the ‘distance’ between pairs of individuals by the statistic: (1-Pearson Correlation), where the Pearson Correlation is computed across the 16 pairs of coefficients. A perfect linear relationship means that the respondents are virtually similar in what appeals to them about geography and what does not. The perfect linear relationship has a Pearson Correlation coefficient of 1.0, and thus the distance between the two respondents is defined as 1-1 or 0. Similarly, when the two respondents show precisely opposite patterns of coefficients, the Pearson Correlation Coefficient is -1, and the distance is 2.0.

The k-means clustering program was instructed to put the 50 respondents into two groups or clusters, and then into three groups or clusters. The computer program [14] divides items (here people) on a purely mathematical basis, according to rules which have nothing to do with meaning of the variables that it uses (here the meaning of the elements who coefficients are used).

When the cluster generates two groups, and then three groups, each respondent is assigned to one of the groups. Table 6 shows the coefficients for the two-cluster solutions, and then the three-cluster solution. The table is sorted by the two-cluster solution.

Table 6: The performance of the elements for Total Panel, Two Mind-Sets, and 3 Mind-Sets, along with the computation of the IDT, Index of Divergent Thought, a measure of ‘how strong the elements performed’ for this clustering solution.

When we look at the two-cluster solution, the patterns becomes far clearer. Here are the very strong elements from the two clusters.

Mind-Set 1 of 2 – focus on the physical world

Physical world topics like natural disasters attract interest through fascination, while human-environment interactions like sustainability challenges appeal through relevance and urgency.

Topics about the physical world intrigue with their unpredictable and powerful forces, while human-environment interactions captivate with their complexity and interdependence.

Physical world subjects draw interest by highlighting the Earth’s forces, whereas human-environment interactions fascinate by exploring the interconnectedness of our actions.

Natural disasters intrigue with their devastating impact, while sustainability challenges inspire us to rethink and protect our environment.

Mind-Set 2 of 2 – Geography helps understand current events and a focus on the methods

Geography enables the exploration of the relationship between resources and economic development in different regions.

Geography provides context for understanding the location and significance of current events.

Geography offers understanding of regional disparities in educational, healthcare, and infrastructure development.

Geography studies regions prone to political instability or unrest.

Understanding scientific methodologies in geography involves exploring techniques for data collection, analysis, and interpretation.

Geographical research employs statistical analysis and data visualization tools to identify trends and patterns.

Geographical research also involves qualitative methods like interviews, surveys, and case studies for a comprehensive understanding of different phenomena.

When we move to three mind-sets, we reduce the parsimony by adding an additional mind-set, but we also increase the interpretability of the mind-sets.

Mind-Set 1 of 3: Focus on the human-environment interaction’

Physical world topics like natural disasters attract interest through fascination, while human-environment interactions like sustainability challenges appeal through relevance and urgency.

Natural disasters intrigue with their devastating impact, while sustainability challenges inspire us to rethink and protect our environment.

Geography unveils the relationship between human behavior and environmental conditions.

Geography helps explain the distribution of resources and their influence on economic development.

Topics about the physical world intrigue with their unpredictable and powerful forces, while human-environment interactions captivate with their complexity and interdependence.

Geography provides insights into the formation and evolution of cities.

Geography illuminates the spatial patterns of human migration and cultural diffusion.

Physical world subjects draw interest by highlighting the Earth’s forces, whereas human-environment interactions fascinate by exploring the interconnectedness of our actions.

Geographical research also involves qualitative methods like interviews, surveys, and case studies for a comprehensive understanding of different phenomena.

Geography studies regions prone to political instability or unrest.

Mind-Set 2 of 3: Focus on geography and society

Geography enables the exploration of the relationship between resources and economic development in different regions.

Geography offers understanding of regional disparities in educational, healthcare, and infrastructure development.

Geography provides context for understanding the location and significance of current events.

Geography studies regions prone to political instability or unrest.

Understanding scientific methodologies in geography involves exploring techniques for data collection, analysis, and interpretation.

Geographical research employs statistical analysis and data visualization tools to identify trends and patterns.

Geographical research also involves qualitative methods like interviews, surveys, and case studies for a comprehensive understanding of different phenomena.

Mind-Set 3 of 3: Focus on research

Geographical research employs statistical analysis and data visualization tools to identify trends and patterns.

Physical world topics like natural disasters attract interest through fascination, while human-environment interactions like sustainability challenges appeal through relevance and urgency.

Understanding scientific methodologies in geography involves exploring techniques for data collection, analysis, and interpretation.

Field surveys, observations, and experiments are conducted to gather primary data for geographical research.

Geographical research also involves qualitative methods like interviews, surveys, and case studies for a comprehensive understanding of different phenomena.

The ability of clustering to pull out these easy-to-interpret groups is worth noting, if only for the fact that to many people participating in these Mind-Genomics studies, the results seem to emerge from guessing.

Table 6: How the elements drive interest when the respondents are clustered or segmented into two, and then into three mind-sets by a statistical algorithm which does not take ‘meaning’ of the elements into account.

At the bottom of Table 6 appear a series of numbers, reflecting computations. These computations are done on the coefficients for the 16 elements, appearing in Table 6 The computations work on the squares of all elements with coefficients of 1 or higher. The computation treats the dataset as if it comprised three groups of 50 respondents each, looks at the sum of squares of coefficients for each column, and weights that column-based sum of squares by the relative proportion of respondents, assuming a set of 3×50 or 150 respondents. The bottom two rows show the weighted sum of squares, and then the square root of the weighted sum of squares.

The final number is the IDT, Index of Divergent Thought. The IDT can only get higher when the elements have high coefficients. The elements can only have high coefficients when all the respondents are thinking in the same way. Thus, the IDT is a measure of the goodness of thinking. The value for the square root is 83. Previous studies suggest that this value of 83 is ‘respectable’ but not exceptionally high. The IDT ends up being a way to measure performance, and the increase of performance over time and practice.

Once the mind-sets are developed, the Mind Genomics program finishes the analysis by summarizing the strong performing elements for each mind-set on a set of queries sent to the embedded AI program. The program answers the queries using only those elements which generated coefficients of +21 or higher. Table 7 presents the AI summarization for the two-mind-set solution.

Table 7: AI summarization of the strong performing elements for the two-mind-set solution

 

Finding These Mind-sets in the Population at Large

The development of knowledge for Mind Genomics is done rapidly, inexpensively, and with easy to reach samples of respondents. Often as few as 50 respondents is needed. Once the knowledge is obtained, however, how can one broaden the impact of the discovery? It is not reasonable to do the Mind Genomics experiment again and again. The knowledge is already in one’s hands. What is missing is the use of the knowledge to mind-type millions of people, prospective readers of a magazine on geography, or a product related to geography. How then can one assign a new person to one of the two mind-sets, with the objective of interacting with this new person, this prospect, on matters relevant to the study?

The idea of an identifying tool, a typing tool, is not new. Typing tools are important in the world of commerce to assign people to relevant groups. Typing tools are used in many applications to assign people to the relevant group. The chances for success working with people is assumed to be higher when the correct messages are sent to the people, messages which resonate with the mind-set to which the person belongs.

Figure 4 shows a mind-set assignment tool, based upon the strong performing elements. The approach, detailed in previous publications [15,16] enables the researcher to identify actual phrases from the study, develop a simple two-point scale (Yes/No), apply that scale to the six questions taken from the study, and then with reasonable confidences (around 70% or higher) assign new people to one of the two mind-sets. Panel A shows the introductory information about the respondent. Each item is optional, but important for practical application. Panel B shows the actual six questions, which are randomized across respondents. It is the pattern of responses to the six questions which drives the assignment. It is with the assignment or typing tool that the Mind Genomics moves from an efficient creator or knowledge to a facilitating mechanism which enhances communication.

Figure 4: The mind-set assigner. Panel A shows the background questions. Panel B shows the six questions, patterns of reaction to which assign a person to one of the two mind-sets.

Discussion and Conclusion

The origin of this paper comes from the opportunity to understand a topic more deeply, the specific topic being geography. Rather than focusing on the substantive issues involved in geography, the paper considers the topic from the point of view of the mind of people. The paper opens up the possibility of incorporating a new viewpoint into studies, the topic-related worldview of the researchers who do the studies. Science is presumed to be objective. Mind Genomics applied to the topic enables a deeper understanding of the research for this very reason.

With a tool such as Mind Genomics in the hands of ordinary people, it is tempting to think what this study about a magazine, or a journal could be. The study as run here comprises only 16 elements, chosen from a great number, and chosen to be run in the United States. One could only imagine the amount of learning one could get about the magazine or journal if one were to explore each of the possible topics in depth, topics such as geography itself, or the relation of geography to economics, geography to social behavior, geography to health, and so forth. Furthermore, the ability to run these studies in different countries, also inexpensively, means that it is well within the realm of possibility to expand this exploration to many different countries. Finally, the ability to run these studies again and again means that the studies can be done for years, as used as a measure of the change of people’s thinking over time.

The second issue is the nature of the respondents. The respondents were chosen virtually at random. The study might have used ordinary people, or just as easily used professionals, or even those involved in geography in one or another fashion. The respondents could have been teachers or geography vs. parents, and so forth. One could put all of the different types of respondents together into one group, done the clustering to generate segments, and determine the degree to which segmentation falls along lines demarcated by one’s profession versus lines demarcated by one’s interest, independent of profession, especially independent of professions as they cross with the topics of geography. In other words, do there exist more profound mind-sets?

The third and final topic is the coming use of artificial or synthetic respondents rather than real people. The use of AI to generate questions and answers at the start of the study through Idea Coach and the ability of AI to summarize the results in the same way, using strong performing elements as the raw materials, both suggest that AI can play an important role in the Mind Genomics effort. The final challenge is AI as the respondent, complementing the contribution as the designer of research inputs, and the analyst of research output. And, if A I can be programmed to act like respondents, does this final step in the AI evolution of Mind Genomics promise new vistas, especially for the world of students, where studies can be thought up, designed, executed, analyzed, and summarized in a matter of hours. What might be the nature of journals and of magazines when AI can be programmed to reflect a specific type of mind. One might get a sense of this future by seeing what types of elements perform well when the AI is instructed to assume a variety of different ‘personas,’ and in the spirit of Mind Genomics, personas designed with attention to the granularity of detail various aspects.

References

1. Lawson AE (1993) Inductive-deductive versus Hypothetico-deductive reasoning: A reply to Yore. Journal of Research in Science Teaching 30: 613-614.
2. Backman K, Kyngäs HA (1999) Challenges of the grounded theory approach to a novice researcher. Nursing & Health Sciences. [crossref].
3. Ellis D (1993) Modeling the information-seeking patterns of academic researchers: A grounded theory approach. The Library Quarterly 63: 469-486.
4. Pickering S (2016) Keeping geography messy. In Teaching Geography Creatively 212-224, Routledge.
5. Scoffham S (2012) Geography and creativity: Developing joyful and imaginative learners. In: Contemporary Issues in Primary Education, 21-33, Routledge.
6. Scoffham S (2016) Teaching Geography Creatively. Routledge.
7. Levin LA, Langer KM, Clark DG, Colquhoun TA, Callaway JL, et al. (2012) Using Mind Genomics® to identify essential elements of a flower product. Horticulture Science 47: 1658-1665.
8. lollari O, Papajorgji P, Civici A (2019) Using Mind Genomics to advertise tourism in .Albania. In: 1st UNICART, Interdisciplinary International Conference on ‘Tourism Management and Development of Territory, Bari, Italy. 270-287.
9. Moskowitz HR (2012) ‘Mind Henomics’: The experimental, inductive science of the ordinary, and its application to aspects of food and feeding. Physiology & Behavior. [crossref]
10. Moskowitz HR, Gofman A, Beckley J, Ashman H (2006) Founding a new science: Mind Genomics. Journal of Sensory Studies 21: 266-307.
11. Gofman A, Moskowitz H (2010) Isomorphic permuted experimental designs and their application in conjoint analysis. Journal of Sensory Studies 25: 127-145.
12. Stevens SS (1966) Personal communication to author Howard Moskowitz.
13. Likas A, Vlassis N, Verbeek J.J (2003) The global k-means clustering algorithm. Pattern Recognition 36: 451-461.
14. Wilkinson L (2010) Systat. Wiley Interdisciplinary Reviews: Computational Statistics, 2: 256-257.
15. Moskowitz H, Gere H (2020) Selling to the mind of the insurance prospect: A mind genomics cartography of insurance for home project contracts. Edelweiss Psy Open Access 4: 22-28.
16. Porretta S, Gere A, Radványi D, Moskowitz H (2019) Mind Genomics (Conjoint Analysis): The new concept research in the analysis of consumer behaviour and choice. Trends in Food Science & Technology 84: 29-33.

DOI: 10.31038/GEMS.2023566

Abstract

In the context of global mobility, the traditional research framework based on a fixed location perspective is facing the need for adaptive innovation and change. In order to deepen the understanding of the evolution of the Anthropocene stratigraphic record and surface landscapes, this paper argues that the nexus approach and the systematic integration of spatial dimensions should be combined in the context of the Anthropocene to form a cross-regional nexus approach, to better cope with the complex issues and global challenges of the Anthropocene.

Keywords

Nexus approach, Cross-regional nexus perspective, Telecoupling, Metacoupling, Spatially dimensional systems integration, Anthropocene

Human beings are ever extending their activity space largely thanks to the technology progress in energy utilization, transport and information communication. The trajectory of human development since the 1760s indicated that the increasing enhancement of regional linkages and connectivity, and the growing higher human and resource mobility have become the prevailing trend and dominated pattern over the world. In these transformative changes, cities play a leading role as engine of the world technology advancement and the regional center for global production and consumption. The increasing connectivity and mobility have also gradually transformed the rural society based on local ecosystem services into the urban society based on non-ecosystem services and long-distance resource redistribution, and have profoundly changed the way humans interacting with nature. One of the most important changes in human-nature interrelationships, especially since the middle of the twentieth century, is the increasing interaction of coupled human and natural systems (CHANS) over long distances [1], and another is the fact that humans are influencing and modifying ecosystems to a greater extent than at any other time in history [2]. Over the past 70 years, human activities have greatly accelerated erosion and weathering on land; greenhouse gas emissions from agricultural, industrial and consumer activities are changing at an unprecedented rate, and the resulting fluctuations in the carbon cycle and climate change are threatening biodiversity and human survival. The changes taking place in the oceans are equally worrisome, and in addition to the widespread concerns about sea-level rise, ocean acidification and marine litter pollution, the recent release of nuclear-contaminated water from Fukushima, which was not sampled by scientists from neighboring countries, has created new uncertainties for marine ecosystems and human survival.

Given that these challenges cross the boundaries of culture, social governance and ecosystems, there is a need to re-examine human-environment interactions from a new perspective. The “Anthropocene” is a geological concept based on the fact that human activities have had a global impact on climate and ecosystems. Since the introduction of the concept of the Anthropocene [3], the field has attracted a wide range of research interest, with recent attempts to establish an Anthropocene gold spike profile being particularly noteworthy [4]. It is worth noting that the anthropogenic factors driving the evolution of stratigraphic records and surface landscapes in Anthropocene have transcended the boundaries of specific regions in the context of evolving globalization and continuing urbanization, making the traditional research framework in the field of Earth sciences based on a fixed locational perspective is facing the need for adaptive innovation and change. In the last decade or so, a series of illuminating research work has been carried out in the nexus approach and spatially dimensional systems integration in order to better understand the interactions between coupled human and natural systems over distances and their resulting socio-economic and environmental impacts across regions. Hoff first introduced the concept of the water-energy-food (WEF) nexus at the Bonn Conference in 2011 [5] to better deals with the challenges posed by global changes from a multi-sectoral perspective. Subsequently, the nexus approach aroused widespread academic interest. In the same year, Liu Jianguo and others proposed the comprehensive concept of Telecoupling at the symposium on “Telecoupling of Human and Natural Systems” at the meeting of the American Association for the Advancement of Science [6]. This theoretical framework for describing socio-economic and environmental interactions at a distance between coupled human and natural systems was elaborated in the later paper “Framing sustainability in a telecoupled world” [7].

The theoretical idea reflected in telecoupling has attracted widespread academic attention, and a series of related theoretical models have emerged in the global academic community in the same period of time that the theory was put forward and thereafter, typically representing “interregional sustainability” [8,9], urban land teleconnections [10,11], local- and tele-coupling [12], metacoupling [13], as well as the Coupled Human and Natural Cube [14]. Among them, metacoupling is a natural extension to the study of telecoupling, and the theoretical framework is actually an integration of intracoupling, pericoupling and telecoupling [13]. It takes into account human-nature interactions within a given system and across spatial distances by incorporating all target coupled systems from near to far, essentially integrating the study of “flow space” and place space [15], and thus not only overcoming the inherent limitations of single-system studies in the classical propositions of human-nature relations, but also compensating for some of the shortcomings of the telecoupling framework [16]. Although many advances have been made in research on the Anthropocene, the nexus approach, and spatial dimensional systems integration, respectively, the combination of the three remains little discussed in the literature. In the context of the Anthropocene, the nexus approach extended to multiple regions – i.e., cross-regional nexus analysis can deepen the understanding of the evolution of the Anthropocene stratigraphic record and surface landscapes. For example, in the framework of the “water-land-food-energy” nexus, water is needed for irrigation and is consumed by energy production that provides electricity for irrigation, while the trend towards the energization of food creates tensions between bioenergy and food production over land resources [17].

These problems of factor clamping and conflict arising from the competitive use of resources affect the food security of distal regions, such as urban areas, through the food system. Conversely, in the process of urbanization, there is also competitive use of soil and water resources in urban areas between different uses. For example, more soil and water resources for urban development or ecological construction means less soil and water resources that can be used for agricultural production. The result is naturally a reduction in the self-sufficiency rate of food in urban areas and a shift in the resource and environmental pressure on food supply to other regions, thus creating a process of cross-regional nexus between natural and human elements [18] (Figure 1). In conclusion, the cross-regional nexus perspective in Anthropocene provides a concrete entry point for revealing the implicit connection and indirect feedback between human beings and the environment in the “flow space”, which is of great significance to the study of human-earth relationship network system under the perspective of metacoupling, and its policy application is of great practical value in solving the many challenges of the Anthropocene and promoting sustainable development. Thus, theoretical and empirical research on cross-regional nexus in Anthropocene deserves the attention of scholars.

Figure 1: A typical case of cross regional nexus – urban food system

Authorship Contribution Statement

Enpu Ma: Conceptualization, Writing original draft, Funding acquisition. Liuwen Liao: Conceptualization, Funding acquisition. Yiwen Ji and Sen Yu: Discussing, Drawing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this original manuscript.

Acknowledgement

This research was funded by the National Natural Science Foundation of China (Grant No. 42101267 and 42101198).

References

1. Liu J, Dietz T, Carpenter SR, et al. (2007) Coupled human and natural systems. AMBIO 36: 593-596.
2. Millennium Ecosystem Assessment Panel. Ecosystem and human well-being: Synthesis. (2005) Washington DC, USA: Island Press.
3. Crutzen P (2002) Geology of mankind. Nature 415.
   
4. Colin NW, Jan Z, Colin S, Ian JF, Neil LR, et al. (2018) Global Boundary Stratotype Section and Point (GSSP) for the Anthropocene Series: Where and how to look for potential candidates. Earth-Science Reviews 178: 379-429.
5. Hoff H (2011) Understanding the Nexus. Background Paper for the Bonn2011 Conference: The Water, Energy and Food Security Nexus. Stockholm Environment Institute, Stockholm.
6. Liu J, McConnell W, Baerwald T, et al. (2011) Symposium on “Telecoupling of Human and Natural Systems” at the meeting of the American Association for the Advancement of Science.
7. Liu J, Hull V, Batistella M, Mateus B, Ruth D, Thomas D, et al. (2013) Framing sustainability in a telecoupled world. Ecology and Society 18.
   
8. Kissinger M, Rees WE (2010) An interregional ecological approach for modelling sustainability in a globalizing world: Reviewing existing approaches and emerging directions. Ecological Modelling 221: 2615-2623.
9. Kissinger M, Rees WE, Timmer V (2011) Interregional sustainability: Governance and policy in an ecologically interdependent world. Environmental Science & Policy 14: 965-976.
10. Seto KC, Reenberg A, Boone CG, Michail F, Dagmar H, Tobias L, et al. (2012) Urban land teleconnections and sustainability. Proceedings of the National Academy of Sciences of the United States of America. 109: 7687-7692.
11. Guneralp B, Seto KC, Ramachandran M (2013) Evidence of urban land teleconnections and impacts on hinterlands. Current Opinion in Environmental Sustainability 5: 445-451.
12. Fang CL, Zhou CH, Gu CL, Chen L, et al. (2016) Theoretical analysis of interactive coupled effects between urbanization and eco-environment in mega-urban agglomerations. Acta Geographica Sinica, 71: 531-550.
13. Liu J (2017) Integration across a metacoupled world. Ecology and Society 22.
    
14. Liu HM, Fang CL, Li YH (2019) The Coupled Human and Natural Cube: A conceptual framework for analyzing urbanization and eco-environment interactions. Acta Geographica Sinica 74: 1-19.
15. Ma EP, Ye WY, Liao LW, Cai JM, et al. (2022) Human-land coupling enlightenment and driving forces of urban food system evolution: A case study of Beijing food system. Journal of Natural Resources 37: 2617-2635.
16. Ma EP, Cai JM, Han Y, Liao LW, Lin J, et al. (2020) Research progress and prospect of telecoupling of Human-Earth system. Progress in Geography 39: 310-326.
17. Food and Agriculture Organization of the United Nations. Bioenergy and food security (2008). China Agricultural Information 6: 4-6.
18. Ma EP, Cai JM, Guo H, Lin J, Liao LW, et al. (2021) Theoretical framework and research priorities on food system couplings in an urbanization context. Acta Geographica Sinica 76: 2343-2359.

DOI: 10.31038/GEMS.2023563

Abstract

Glaciological studies have been conducted on several snow patches in Japan since 1910^th,, including a recent report with the conclusion that some perennial snow patches in the Hida Mountains in north-central Japan are identified as glaciers. The identification was supported by all Japanese glaciologists at an open symposium organized by the Data Center for Glacier Research of the Japanese Society of Snow and Ice, recognizing that the snow patches have considerable amount of ice mass in the lower part of the snow patch, which shows the evidence of flow downwards along the valley. The international cryosphere-related organizations, however define the glacier as “stagnant or flowing ice mass”, and the international classification of glaciers published by the International Commission of Snow and Ice (ICSI) shows the necessity of flow only for some classes of glaciers such as the valley glacier, outlet glacier etc., and thus, the flowing is found not principally necessary to identify some groups of glaciers. Examining comparatively the pilot study in the Himalayan region given by the ICSI, several snow patches with ice masses in Japan are considered to belong the valley glacier, mountain glacier, or glacieret. However, the ICSI classification above includes vagueness on flow, size and thickness of the ice mass, distinction among valley glacier, mountain glacier, glacieret and snow patch. Those points should be examined and clarified for the better definition and classification of snow and ice masses on the earth.

Introduction

There is a variety of snow and ice masses on the globe. Some of them have been regarded as the glacier. However, the definition and classification of the glacier is not clearly understood even in researchers of related sciences. In Figures 1-3 some typical glaciers, which should be understood generally as the valley glacier, mountain glacier, and ice sheet with an outlet glacier are given.

Figure 1: Aretsch Glacier in the Swiss Alps, view of the middle reaches. Far above, uppermost cirques are seen (Photo provided by R. Naruse, taken in July 2018).

Figure 2: West side view of Arakam tse, Everest area, with typical mountain glaciers on the west slope. A small valley glacier is seen on southeast of the peak and a large Nojumba Glacier is seen on the bottom of the picture. (Photo taken in October 2007).

Figure 3: Antarctic ice sheet in eastern Queen Maud Land and Skallen outlet glacier, February. 1970

In 2012, 3 perennial snow patches in Tateyama-Tsurugi range of Hida Mountains in north-central Japan were identified as active glaciers by Fukui and Iida (2012) [1]. About one hundred glaciologists in Japan assembled and discussed on the report at an open symposium [Perennial snow patches and glaciers in Japan – past studies and future researches organized by the Data Center for Glacier Research of the Japanese Society of Snow and Ice (JSSI) [2]. The summary of the symposium is that the participants of the symposium do not object to call these 3 snow patches as the glacier.

The present author has had some experiences of observing glaciers worldwide as a geologist and a mountaineer since the 1960s, and has been interested in the active glaciers reported by Fukui and Iida. It is found that there are some vagueness to be clarified on the definition and classification of glaciers. The present author proposed in 2019 (Yoshida, 2019-2020) [3,4] to call the above kind of snow patches as the “snow patch glacier” (abbreviated as SPG from here onwards in this paper). Recent researches in Japan on SPGs are considered useful to contribute the understanding of similar SPGs, which are distributed widely in some areas of the earth.

Glaciological Research on Perennial Snow Patches in Japan

The beginning of glaciological studies on perennial snow patches in Japan was the report by Ohzeki (1917) [5] on a perennial snow patch (designated as PSP below onwards) in Mt. Gassan in Northeast Japan. Ohzeki showed the shape, structure, and size of the ice mass within a large perennial snow patch known as “Ohyukijiro” (big snow castle, in Japanese meaning), and pointed out a possibility that the snow patch is similar with the glacieret distributed at several places in European Alps.

Since then several glaciological studies on snow patches in Japan including the report of identification of glaciers or glacierets were conducted [1,6-31] in mountains in Hokkaido, Northeast Japan and north central Japan, etc.

After around 1970, a systematic study project of snow patches in Japan started led by Prof. K. Higuchi of Nagoya University. The study project was related with the International Hydrological Decade (IHD, 1963-1974) and Intergovernmental Hydrological Programme (IHP, 1975-) that aimed to clarify the actual situation of water resources and to contribute to the examination of their relationship with climatic changes. Due to the studies of the above Japanese project, the outline of perennial snow patches distributed throughout Japan was clarified and detailed studies of several snow patches were also conducted. All these studies contributed to the IHD-IHP projects for the formation of the world glacier inventory.

The glaciological researches in Japan had been influenced by the general idea of Japanese geomorphologists that there should be no glaciers in Japan referring the suggestion of Hoshiai and Kobayashi (1957) [32] that the equilibrium line altitude (snow line) in Japan is 4000 meters, which is nearly one thousand meters higher than the Japanese highest mountain range. Several glaciological researches in Japan, however, have been conducted even after 1957. It is first because field evidence of glaciologic characteristics of some snow patches were noticed, and second because the idea that the snow line might change locally due to differences in local climate and topography, which affect in large differences in the amount of snow accumulation and melting. The above idea was further clearly pointed out recently by Ono [33].

In 2012, Fukui and Iida [1] reported 3 PSPs in Tateyama-Tsurugi range in the Hida Mountains (Japanese North Alps) in north-central Japan are identified as active glaciers. These PSPs are several hundred to over one km in length, have a thick (25 – 45 meters thick) ice bodies in the lower horizons, and are flowing downwards some meters per year. They showed mesoscopic and microscopic inner structure of the ice mass and figured out the mechanism of ductile flow of the mass. Fukui and Iida (2012) [1] pointed out that the above signatures are conformable with the definition of glaciers.

An open symposium [Perennial snow patches and glaciers in Japan- past studies and future researches] was held in 2012 at the Data Center for Glacier Research of JSSI to discuss on the report of Fukui and Iida (2012) [1], and about 100 members of the society assembled.

During the symposium, several lectures and discussions were presented in relation to the report of Fukui and Iida (2012) [1]. At the end of the symposium, the chairman summarized that all the participants did not deny that 3 PSPs reported by Fukui and Iida could be called as the glacier, but one PSP has to be further examined, because the flow of the PSP is not clear [34]. In the symposium, however, there appeared to have been no discussion on the criteria of definition and classification of glaciers, and further, no discussion referring those by international organizations such as the International Glaciological Society (IGS), ICSI, etc. Almost all participants of the symposium appeared to have had the understanding on the definition of glaciers as that given in the so far published dictionaries of JSSI [35].

Fukui and others [29] and Arie and others (2019) [30] further reported additional 4 PSPs in the Hida Mountains with the conclusion that they also have signatures to be identified as glaciers. Similar PSPs in the Hida Mountains are under the survey by the same authors and will shortly be reported. The distribution of these snow patches are shown on topographic maps in Figures 4-6. In these maps, all large snow patches shown on the 1:25000 topographic map of the Geospacial Information Authority of Japan (2018) [36] are delineated by the present author. Snow patches, which were so far reported [1,29-31] that they have or should have an ice mass are shown as SPG. Field photos of some of the above SPGs are given in Figures 7-9. A detailed topographic sketch and a cross section of one of the SPGs are shown in Figure 10a and 10b, citing after Fukui et al. [29].

Figure 4: Distribution of SPGs and a PSP surrounding Mt. Tateyama in the northwestern Hida Mountains. A: Ikenotani SPG, B: Komado SPG, C: San-no-mado SPG, D: Tsurugisawa PSP, E: Bessanzawa PSP, F: Hamaguriyuki SPG, G: Kuranosukesawa SPG, H: Gozensawa SPG.

Figure 5: Distribution of SPGs and PSPs on the eastern slope of Mts.Shiroumadake-Karamatsudake in theNortheastern Hida Mountains. A: Hakubasawa migimata PSP, B: Hakubasawa SPG, C: Daisekkei PSP, D: Shakushizawa SPG, E: Tengusawa PSP, F: Kaerazusawa SPG, G: Karamatsuzawa SPG.

Figure 6: Distribution of SPGs and PSPs in eastern slopes of Mts.Karamatsudake-Kashimayarigatake, northeastern Hida Mountains.A: Karamatsuzawa SPG, B: Ohgurozawa-migimatamigi PSP, C.

Figure 7: San-no-mado SPG and Komado SPG on the eastern slope of Mt. Tsurugidake (Aerial Photo Provided by J. Akabane)

Figure 8: Karamatsuzawa SPG and Kaerazuawa SPG on the eastern slope of the north ridge of Mt.

Figure 9: Kakunezato SPG on the north- east slope of Mt. Kashimayarigatake, northeastern Hida Mountains

Figure 10a: Horizontal sketch of Kakunezato SPG in October 2015 (after Fukui et al., 2018)

Figure 10b: Cross section of the Kakunezato SPG (Referred to Fukui et al., 2018)

Commentary on the Definition and Classification of Glaciers

There is so far no definition or criteria of recognition of glaciers officially decided by international as well as Japanese academic societies of related sciences. Merrian-Webster Dictionary [37] gives the definition of the glacier as “large ice body flowing slowly on the slope or along a valley, or large ice body on a large gentle land and flows outside from the land”, and a Japanese dictionary of earth science [38] describes “ice body existing on the surface of the earth and flowing continuously”. These descriptions may give a general understating of a glacier to people. However, we have to examine the correct definition, which is acceptable by scientists of related fields from the cryospheric academic point of view.

In the cryospheric world of Japan, explanations given by Ageta [35] and Shiraiwa [39] appeared in dictionaries of snow and ice convened by JSSI could generally be accepted by Japanese researchers of related fields, although Naruse [40], citing the definition of Flint [41], gives the explanation of a glacier as 1) large mass composed of snow and ice, and 2) existing on land and flows presently or flowed in the past.

The definition given by Flint [41] in his book Glacial and Quaternary Geology, “a large mass of snow and ice accumulated naturally and is mostly existing on the land, and is flowing presently or flowed in the past” is often referred to internationally. Reisco [42] pointed out that the International Association of Cryospheric Sciences and Intergovernmental Panel on Climate Change (IPCC) also show the similar definition.

In contrast, the IGS [43] has declared the agreement upon the Standard Practice Guide used by the Argentine Institute of Nivology, Glaciology and Environmental Sciences for making the National Glacier Inventory of Argentine, ① “a snow and ice mass with the dimension of less than 0.01 km² is not included in the inventory, and ② definition of glaciers as “glacier is a perennial ice mass either stable or flowing” that appeared in the Argentinean Glacial Law. Raup [44] showed a candidate definition of the glacier related with the GLIMS (Global Land Ice Measurements from Space) Project of NSIDC (National Snow and Ice Data Center of US) [45] as “an ice mass existing over several years and has the dimension of more than 0.01km² after the melting in summer”, and no description on the flow of the ice mass is given. The amount and thickness of ice mass within the snow and ice mass are considered to be critical for the recognition of a glacier, however, are not referred to in any descriptions of definition of glaciers so far disclosed. This point will be discussed in a later section.

Thus, there are two different ideas on the definition of glaciers regarding its flow signature, i.e., either “an ice mass stable or flowing” or “an ice mass flowing presently or flowed in the past”. The author prefers the former, because it is often almost impossible to clarify either an ice mass is a remnant of a past glacier or not. Since an ice mass has existed in a snow patch during this warm interglacial stage, it is possible that the ice mass of a past glacier of the glacial epoch have not disappeared and existed continuously throughout the interglacial stage. In such a case, it does not matter whether ice crystals forming the mass might have renewed time to time and no old crystals formed during the glacier epoch might have survived or not.

The principal classification of glaciers is based on temperature conditions (e.g., Ahlmann [46]; Journal of Glaciology, 1957) [47]. However, for snow and ice masses occurring in a restricted area, the above classification has almost no importance and instead, classifications based on the topographic characteristics are generally adopted (e.g., United Stage Geological Survey, 2004).

The International Committee on Snow and Ice (ICSI) planned under the International Hydrologic Decade (IHD, 1965-1974), with the objective to outline all snow and ice masses of the earth, and disclosed in 1967 a table of classification of glaciers (Table 1, UNESCO-IASH, 1970, described as ICSI table below) as the standard practice guide for making a world catalogue of snow and ice masses (the world glacier inventory) on the surface and sub-ground of the earth. The program of making the world inventory of glaciers has been succeeded to the IHP (1975), and in 1995 NSIDC started the renewal of the world glacier inventory project (GLIPS Project) [48], and the first report was delivered in 2014. In all the above international projects, the ICSI table has actively been used.

Table 1: Classification of snow and ice masses (UNESCO/IASH, 1970)

	Digit 1 Primary classification	Digit 1 Form	Digit 3 Frontal characteristics	Digit 4 Longitudinal profile	Digit 5 Major source of nourishment	Digit 6 Activity of tongue
0	Uncertain or misc.	Uncertain or misc.	Normal or misc.	Uncertain or misc.	Uncertain or misc.	Uncertain or misc.
1	Continental ice sheet	Compound basins	Piedmont	Even; regular	Snow and/or drift snow.	Marked retreat
2	Ice-field	Compound basin	Expanded foot	Hanging	Avalanche ice and/ or avalanche snow	Slight retreat
3	Ice cap	Simeple basin	Lobed	Cascading	Superimposed ice	Stationary
4	Outlet glacier	Cirque	Calving	Ice-fall		Slight advance
5	Valley glacier	Niche	Coalescing, non-contributing	Interrupted		Marked advance
6	Mountain glacier	Crater				Possible surge
7	Glaieiret & Snow field	Ice apron				Known surge
8	Ice shelf	Group of small unites				Oscilating
9	Rock glacier	Remnant

The ICSI table has been used widely in the snow and ice academic world (e.g., Rau et al., [49]; JSSI, 2019). This table, however, was principally prepared to include all masses of snow and ice of the earth. That means that it includes not only all kinds of glaciers but also snow and ice masses such as snowfield, snow patch, ice shelf, etc., which are obviously not glaciers, and is thus practically the table for perennial snow and ice masses. The GLIMS Project disclosed a table principally same as the ICSI table, however, is supplemented by detailed explanations for each kind of snow and ice masses, and thus enables to give common and detailed classification of snow and ice masses in the world.

The constitution of the ICSI table, as shown on the Table 1 is briefly explained below. The table consists of 6 digits. In digit 1, all snow and ice masses on the earth are classified into 9 primary classes. Digits 2 to digit 6 are composed of characteristics of frontal shape, longitudinal profile, major source of nourishment, and activity of tongue of the 9 classes above, such as existence/absence, strength, size, etc. in 4 to 10 levels. Simple sketches and explanations are given for each class. Some explanations of the classes, which appear to be related with PSPs and SPGs of Japan mentioned above, are given below extracted from the explanations attached with the ICSI table.

Valley glacier: The glaciers of this class flow in the valley, and have clear nourishment area. Both side slopes of the valley are generally not covered by ice.

Mountain glacier: This class includes cirque, niche, and crater glaciers as major kinds, and apron, hanging and group glaciers also belong the same class. The topography of these glaciers does not show the development into the valley shape, however, the distinction between the valley and mountain glaciers are sometimes not easy.

Glacieret and snowfield: This class includes small masses of snow and ice with a variety of shapes that resist existing over 2 years. They are formed by drift, avalanche, or unusual precipitation on depressions, valley floor, or sun or window shades. They generally have no indication of flow. It is difficult to clearly distinguish between the glacieret and snow patch. Nourishment and ablation areas are sometimes not identified.

The explanation of ICSI table given on the glacieret appears difficult to understand because the table explains altogether with glacieret and snowfield. The Cryosphere Glossary given by NSIDC, which is considered to be intimately related with the GLIMS Project describes the glacieret as “a very small glacier”. To examine altogether the above, we understand that the glacieret and snowfield given in the ICSI table is composed of two groups, i.e., the glacieret, which is a very small glacier, and the snowfield, which includes snowfield, snow patch, and snow deposit. There is no clear evidence of flow in both of the above groups.

The ICSI table is, as mentioned above, completely different from proper dictionaries cited above. It classifies all kinds of snow and ice masses and gives detailed explanations. Further the manual of glacier classification by GLIMS cited above gives further detailed explanations on many kinds of glaciers of the table. In this classification, all snow and ice masses except snowfield, snow patch, ice field and ice field are identified as glaciers. That means that a snow and ice mass which is classed into either one of the classes 1, 3, 4, 5, 6, 9, and glacieret of 7 of the ICSI table is identified internationally as a class of the glaciers. In conclusion, the ICSI table is further useful and clear than ordinary dictionaries of related science world to define either a snow and ice mass is the glacier or not, and in what class of the glacier it belongs.

To avoid confusion on the definition and classification of glaciers, the usage of the international definition of glaciers (IGS, 2017) [43] and ICSI table is preferable. Actually in Japan, JSSI discloses the same table (with Japanese translation by G. Wakahama – [50]) as the model for the world inventory of snow and ice masses.

In the above ICSI table including detailed explanations of GLIMS, indexes on the size of the snow and ice mass and speed of flow are not shown, and thus, a variety of snow and ice masses with no flow and with a variety of sizes are included. There is no description on the flow signature for the mountain glacier, and for the glacieret, a description “no obvious flow is recognized” is given. For other glaciers, for which the flow of the mass is mentioned, there is no constraint on the speed and mechanism of the flow and the inner structure of the mass, and even on what part of the snow and ice mass flows is not given. Further for all snow and ice masses, no description on the past or future flow is given.

It may be reasonable that the description on the flow is not necessary in the ICSI table, since the table was principally made with the objective (UNESCO/IAASH, 1970) [51] of clarifying the amount and the income and outcome of snow and ice of the earth. However, since the table is to be used now, to cover all snow and ice masses of the earth and to show their major signatures apart from its original objective, existence/absence of and speed of the flow, size, and amount of ice mass of a snow and ice mass are expected to be added to the ICSI table as the 7^th-9^th digits.

UNESCO/IASH (1970) [51] disclosed examples of usage of the ICSI table for making the world glacier inventory for snow and ice masses, in the North Polar region, Rocky Mountains and Himalaya in pilot field studies, showed general examples of field evidence of dimensions and ice thicknesses for some classes of glaciers as follows.

Valley glaciers: 1-100 km² in dimension and 30-120 m in ice thickness.

Mountain glaciers: 0-20 km² in dimension and 20-120 m in ice thickness.

Glacierets: 0-2 km² in dimension and 10-25 m in ice thickness.

As mentioned above, the critical smallest limit for the dimension of a snow and ice mass as 0.01km² is considered to have become used worldwide for making the world inventory (e.g., Paul et al. [52]; IGS, 2017) [43]. The Journal of Glaciology [47] once showed the idea that an ice mass of less than 15m in thickness could not be included in the inventory. However after that, the ice thickness of 15 meters appears not become common for the registration in the inventory. Muller (1970) [53-56] in his pilot study in the Himalaya showed several glaciers with the ice thickness of 8 meters, although the thickness was not practically measured.

Muller [53] in his study of glaciers in the Everest area of eastern Himalaya as the pilot study of UNESCO/IASH (1970) [41], showed many examples of signature (length, width, ice thickness, etc.) of glaciers. Some examples of his description are given in Figures 11, 12 and Table 2. In his study, the valley glaciers develop generally in a large, long and gentle valley and the snow and ice mass is very large, thus, the distinction from mountain glaciers are generally clear. However, there are found some exceptional cases that a short glacier developed on a steep mountain slope is also classed as the valley glacier. Almost all mountain glaciers are classed as the cirque glacier or unclassed, and only 2 glaciers among 121 mountain glaciers are classed as the niche glacier. Only 2 glacierets are shown in the report and they develop on steep mountain slopes and the distinction from the mountain glacier is not clear.

Figure 11: Glaciers surrounding Kyajo Ri, west of Dudhu Koshi, Everest area. Numbers show glacier numbers and locations reported by Muller, 1970.

Figure 12: The western slope of Kyajo Ri, with glaciers (numbered) described by Muller (1970)

Table 2: Description of glaciers in the area of Fig. 10 (surrounding Kyajo Ri) by Muller (1970). The classification numerical numbers are given by Muller following the Table 1.

Glacier No.	Classification	Length km	Dims km2	Ice thick m
NEA45C-55	600011	1	0.29	13
NEA45C-56	600011	1.1	0.36	14
NEA45C-57	602311	1.4	0.52	18
NEA45C-58	600210	0.8	0.06	10
NEA45C-59	640020	0.4	0.06	10
NEA45C-60	640021	1	0.24	20
NEA45C-61	700210	0.9	0.11	10
NEA45C-62	600210	0.5	0.13	12
NEA45C-63	600210	0.2	0.06	10
NEA45C-64	600310	1.1	0.06	10
NEA45C-65	640022	0.4	0.15	12

Referring all the above, SPGs of Japan developed in small valleys and have been identified as glaciers are considered to belong the mountain glacier, glacieret, or valley glacier. Fukui et al. [29] classed all 6 SPGs they reported to belong the mountain glacier, among which 4 glaciers as the niche, and 2 as the cirque glaciers. The author, however, prefers to classify most of those glaciers as the valley glacier, since these SPGs develop in deep and clear valley topography, which are different from niche.

The cirque glacier is considered to be the glacier that contributed principally to form the cirque topography. Two glaciers classed as the cirque glacier by Fukui et al. [29] are not clear either they contributed to the formation of the concerned cirque topography or not, and therefore they could be provisionally classed as the valley glacier or glacieret.

Summary and Discussion

1. The glaciological study on perennial snow patches (PSPs) in Japan has been conducted continuously since around 1917. Since 1970, glaciologic data were summarized by K. Higuchi and others for wide areas in Japan. In 2012, 3 PSPs in the Hida Mountains were identified as the active glacier, and the Data Center for Glacier Research of JSSI had an open symposium to examine the above identification. The symposium positively recognized it, appreciating the existence of distinct flow of the ice mass as the important criterion for the identification of the glacier.
2. The definition and classification of glaciers are examined referring various references. There are two different ideas on the definition of glaciers regarding its flow signature, i.e., either “an ice mass stable or flowing” or “an ice mass flowing presently or flowed in the past”. The author prefers the former, because it is often almost impossible to clarify either an ice mass is a remnant of a past glacier or not. Regarding the dimension, 0.01 km² is also the widely accepted criterion as the smallest limit, and there is no constraint on the dimension and thickness of the ice mass within a snow and ice mass.
3. The definition of a glacier given in common dictionaries gives only a general signature of glaciers and does not include all kinds of glaciers. The glacier classification table by ICSI, in contrast, includes all kinds of glaciers and provides detailed characterization of them, and therefore useful to define, identify and classify all kinds of glaciers and snow and ice masses, although the table includes snow and ice masses such as snowfield and iceshelf, which do not clearly belong the glacier.
4. The international glacier classification table of ICSI has been used widely in international bodies and projects. This table, however, was made with the objective of classifying all kinds of snow and ice masses of the earth, and therefore includes all kinds of glaciers as well as snowfield, snow patch, ice shelf etc., which are obviously not the glacier.
5. Large (larger than 0.01km² in dimension) perennial snow patches in Japan with considerable size of ice mass within it are classified as glacieret, mountain glacier or valley glacier. The distinction among the above three kinds of glaciers, however, includes some uncertainties, and ideas of classification is not definite at present.
6. Judgement criteria to identify a snow and ice mass as the glacier is simple following the international definition. Size of the ice mass within a snow and ice mass, identification of flow, speed of flow, inner structure, and mechanism of flow of the ice mass are important to understand the glaciological signature of the ice mass, but not necessary for the identification of the glacier.
7. The present international classification table of glaciers (ICSI table) includes some uncertainties and needs to include some components in the criterial signatures such as: existence/absence of flow of ice, and size of the ice mass. Further, some more characterizations for the discriminations among valley glacier and mountain glacier, mountain glacier and glacieret, niche glacier, glacieret and snow patch are considered necessary. The present nomenclature of one of the principal classes of the ICSI table, the mountain glacier, is confusing, because it is generally understood as glacier developed in mountains. An appropriate name such as the mountain slope glacier proposed by G. Wakahama (in JSSI, 1970) may be worth consideration. These points are important to clearly classify a snow and ice mass and requested to be examined in the related science world.

Acknowledgement

The present author is indebted to Dr. Renji Naruse of the Glacier and Cryospheric Environment Research Laboratory in Japan, who gave important advices at the beginning of the precursory Japanese manuscript and also provided a photo of the Aretch Glacier, and Mr. Jinyu Akabane and Dr. Chiyuki Narama who provided photos of snow patch glaciers of the Hida Mountains.

Reference

1. Fukui K, Iida H (2012) Identifying active glaciers in Mt. Tateyama and Mt. Tsurugi in the northern Japanese Alps, central Japan (in Japanese)．Seppyo ( Jap. Soc. Snow Ice) 74: 213-222.
2. Japanese Society of Snow and Ice home page， 2018].
3. Yoshida Masaru (2019) Glaciers of Japan and world－from an excursion to view the Kakunezato glacier, 1 and 2 (in Japanese). Chigaku kyoiku to kagaku undoh (Education in Geoscience and Movement for Science) 83: 56-62 and 84: 64-70.
4. Yoshida Masaru (2020) Active glaciers in Japan－Definition and classification of glaciers (in Japanese)*. Chishitugaku zasshi ( Geol. Soc. Japan) 126: 149-156.
5. Ohzeki K (1917) On perennial snow patches in Gassan, Uzenkoku* (in Japanese)． Chishitugaku Zasshi ( Geol. Soc. Jap) 24: 366-373.
6. Imanishi K (1929) On perennial snow of Tsurugizawa*(in Japanese)． Chikyu (The Earth) 11: 267-
7. Sakita R (1931) On an ice mass lying at the valley head of Gozenzawa, Tateyama* (in Japanese)． Chirigaku hyouron ( Rev., Jap.) 7: 734-747．
8. Iozawa T (1959) A note on Kakunezato (in Japanese)．Chiri (The Geography) 4: 96-104.
9. Ogasawara K (1964) Glaciers in the Northern Japanese Alps: a proposal of the distinctive temperate glaciers* (in Japanese). Toyama Daigaku Gakujutu Chosa-dan, (Academic Research Group of Toyama Univ.)* Ed, Pg: 5-34. Kokon Shoin, Tokyo.
10. Yoshida J (1964) Glaciological survey of perennial snow patches in Tateyama* (in Japanese). Toyama Daigaku Gakujutu Chosa-dan, ed. Tateyama no mannenyuki no.
11. Wakahama G, Endo Y, Yamada T, Ushiki H, Naruse R (1968) Studies of firn on Mt. Daisetsu in summer，III (in Japanese). Low Temperature Science, A, 26: 215-229.
12. Higuchi K (1968) A review of the glaciological studies on perennial ice and snow masses in Japan (in Japanese). Seppyo ( Jap. Soc. Snow Ice)30: 195-207.
13. Higuchi K (1979) Collected papers on perennial snow patches in Japan (1) (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 41:1-7．
14. Higuchi K, Watanabe O, Ushiki H, Okudaira H, Ageta Y (1971) Glaciological studies on the perennial snow patches in Tsurugizawa, Part I -Glaciological studies on the perennial snow patches in Tsurugizawa, Part I- (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 32: 129-146．
15. Higuchi K, Ohata T, Watanabe O (1979a) Change in the size of the “Hamaguriyuki” perennial snow patch on Mount Tsurugi, central Japan (in Japanese)．Seppyo ( Jap. Soc. Snow Ice) 41: 77-84.
16. Higuchi K, Wakahama G, Yamada T, Naruse R, et al., (1979b) A summary on the studies of perennial snow patches in Japan (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 41: 181-197．
17. Tsuchiya I. (1974) Survey on glacierets in Chokaizan* (in Japanese). Chiri (The Geography) 19: 51-59．
18. Tsuchiya I (1976a) Kaigata Glacieret in Chokaizan (in Japanese). Chiri (The Geography) 26: 19-28．
19. Tsuchiya I (1976b) Annual deviation and its meteorological analysis of glacierets in Chokaizan* (in Japanese)．Tenki (The Weather) 23: 53-56．
20. Tsuchiya I (1976c) Airborne photogrametrical estimations of extraordinary snow depths in Mts Iide, Gassan and Chokai and a preliminary report n the birth of a small glacier. Seppyo ( Jap. Soc. Snow Ice) 38: 22-31．
21. Tsuchiya I (1977) The Glacio-climatological study on the Kaigata Glacieret, Mt. Chokai (1) Inter annual variation and structure of annual boundary layers (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 39: 201-212．
22. Tsuchiya I (1978a) The glacio-climatological study on the Kaigata Glacieret, Mt. Chokai 2, Instrumental surveys of shape and size and movement measurements (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 40: 1-9．
23. Tsuchiya I (1978b) The Glacio-Climatological Study on the Kaigata Glacieret, Mt. Chokai (3) Thermal analysis in combination with remote sensing and some features as glacier (in Japanese). Seppyo ( Jap. Soc. Snow Ice) 40: 10-21.
24. Tsuchiya I (1999) Nihon no mannenyuki (Perennial snow patches in Japan)* (in Japanese)． Kokon Shoin, Tokyo
25. Matsuda M (1976) Structure of perennial snow patches in Mt. Daisetsu area, Hokkaido (in Japanese). Seppyo 38: 115-126.
26. Yoshida, Minoru, Yamamoto K, Higuchi K, Iida H, et al., (1990) First discovery of fossil ice of 1000-1700 year B.P. in Japan. Glaciol. 36: 258-259.
27. Kawashima K, Noguchi Y, Kobayashi T, Sato M (2000) Glaciological investigations of the Yamunaisawa perennial snow patch by means of shallow core drillings (in Japanese)．Rishiri Kenkyu (Rishiri Research) 19: 79-87.
28. Fukui K, Iida H, (2013) Reply to “Comments on identifying active glaciers in Mt. Tateyama and Mt. Tsurugidake in the northern Japanese Alps, central Japan” by I. Tsuchiya (in Japanese)． Seppyo ( Jap. Soc. Snow Ice) 75: 147-149．]
29. Fukui K, Iida H, Kosaka K (2018) Newly identifying active glaciers in the northern Japanese Alps and their characteristics (in Japanese)． Chirigaku hyouron ( Rev. Jap.) Ser A 91: 43-61．
30. Arie K, Narama C, Fukui K, Iida H, Takahashi K (2019) Ice thickness and flow of the Karamatsuzawa snow patch, northern Hida Mountains. Seppyo ( Jap. Soc of Snow and Ice) 81: 283-295.
31. Yoshida Masaru (2022) Snow patch glaciers surrounding Hakuba village, Nagano Prefecture―possibility of forming a glacier park/geopark (in Japanese)*. Chigaku kyoiku to kagaku undoh (Education in Geoscience and Movement for Science) 89: 25-34.
32. Hoshiai M, Kobayashi K (1957) A theoretical discussion on the so-called “snow line”, with reference to the temperature reduction during the Last Glacial Age in Japan. Japanese Journal of geology and Geography, 28: 61-75.
33. Ono Y (2012) Studies on glacial landforms in Japan between 1960 and 2010: Personal review (in Japanese)． Chigaku Zasshi ( Geogr) 121: 187-214.
34. Shiraiwa T, Naito N, Iida H, Fukui K (2012) Report of the Public Symposium of the Center for the Glacier Research Information: Perennial snow patches and glaciers in Japan-Previous research and a future view* (in Japanese)． Seppyo ( Jap. Soc. Snow Ice) 74: 353-357.
35. Ageta Y (1990) [Glacier] and [Glacier flow], in Japanese Society of Snow and Ice (Ed.), Dictionary of Snow and Ice (in Japanese), Kokon Shoin, Tokyo, p. 100 and p.103.
36. Geospacial Information Authority of Japan (2018) Denshikokudo.
37. Merrian-Webster Dictionary (2018) Definition of glacier.
38. Toya H, Iwata S (1996) Chigaku Jiten, New Edition* (in Japanese), Ed. Commit Chigaku Jiten (Dictionary of Geoscience). Geol. Collab. Jap 1993.
39. Shiraiwa T (2005) Definition, Classification and Movement of Glacier* (in Japanese). Japanese Soc. Snow Ice (superv)*, Dictionary of Snow and Ice*, 277－287．
40. Naruse R (2013) Dictionary of Glacier and Cryosphere, 5^th Edition (in Japanese). Glacier and Cryospheric Environment Research Laboratory HP, 2018.
41. Flint RF (1971) Glacial and Quaternary Geology. Wiley and Sons NY 39-44.
42. Reisco M (2015) Glacier resources law and policy: A case of transboundary management. In: International Water Resources Association (IWRA) HP.
43. International Glaciological Society (2017) Position paper of the International Glaciological Society on standard practices regarding glacier inventories. IGS-HP.
44. Raup B (2006) Proposed GLIMS Definition of “Glacier”.
45. NSIDC (National Snow and Ice Data Center), 2019, Cryosphere Glossary. In: NSIDC HP.
46. Ahlmann H (1935) Contribution to the physics of glaciers. J 86: 97-113.
47. Journal of Glaciology (1957) The Classification of Glaciers. Glaciology 3: 21, 2-7.
48. Pfeffer WT, Arendt AA, Bliss A, Bolch T (2014) The Randolph Glacier Inventory: a globally complete inventory of glaciers. Glaciology 60: 537-552.
49. Rau F, Mauz F, Vogt S, Khalsa SJS, Raup B (2005) Illustrated GLIMS glacier classification manual. GLIMS Regional Center ‘Antarctic Peninsula’, Pg: 1-36.
50. Wakahama G (1978) Translation from “A guide for compilation and assemblage of data for a world inventory, IHD Resolution, 1967” (in Japanese). Soc. Snow & Ice HP.
51. UNESCO-IASH (1970) Technical papers in hydrology, 1. Perennial ice and snow masses, a compilation to the International Hydrological Decade. UNESCO, Paris, Pg: 1-53.
52. Paul F, Barry RG, Cogley JG, Frey H, et al., (2009) Recommendations for the compilation of glacier inventory data from digital sources. Annals of Glaciology 50: 119-126.
53. Müller P (1970) A pilot study for an inventory of the glaciers in the Eastern Himalaya. In: UNESCO-IASH, 1970, Perennial ice and snow masses Pg: 47-59.
54. Fountain AG (2006) Proposed GLIMS definition of “glacier”. In: NSIDC-HP, 2019.
55. USGS (United States Geological Survey) (2017) Types of glaciers. In: Glossary of Glaciers Terminology – text version, USGS Homepage.
56. seppyogakuteki chousa (cryospherical survey of perennial snow patches in Tateyama). In Tateyama no Shizen (Nature of Mt. Tateyama)*, 5－34．Kokon Shoin, Tokyo.