Computer-based data acquisition and visualization systems in fi eld : Results from 12 years of experimentation and future potential

Terry L. Pavlis, Richard Langford, Jose Hurtado, and Laura Serpa Department of Geological Sciences, University of Texas at El Paso, El Paso, Texas 79968, USA

ABSTRACT attributing is limited to critical information tem (GPS) receivers, digital cameras, recording with all objects carrying a special “note” fi eld , compass- Paper-based geologic mapping is now for input of nonstandard information. We devices that log data and position, and laser archaic, and it is essential that suggest that when fi eld GIS systems become rangefi nders allow us to do things that were transition out of paper-based fi eld work and the norm, fi eld geology should enjoy a revo- impossible only a decade ago. This technology embrace new fi eld geographic information lution both in the attitude of the fi eld geolo- allows new approaches to obtaining, organizing, system (GIS) technology. Based on ~12 yr gist toward his or her data and the ability to and distributing data in all fi eld sciences. This of experience with using handheld comput- address problems using the fi eld information. technology will undoubtedly lead to major new ers and a variety of fi eld GIS software, we However, fi eld geologists will need to adjust advances in fi eld geology, and hopefully revital- have developed a working model for using to the changing technology, and many long- ize this foundation of the geosciences. fi eld GIS systems. Currently this system uses established fi eld paradigms should be reeval- In this paper we explore this issue of changing software products from ESRI (Environmen- uated. One example is the rule that all line- fi eld technology based on 12 yr of experience, tal Systems Research Institute, Inc.) (ArcGIS work on geologic needs to be perfected during which we have progressed from using and ArcPad), but the data model could be in the fi eld setting. Our experience suggests computerized fi eld notebooks to using a variety applied to any GIS system. This fi eld data that with modern high-resolution imag- of fi eld GIS systems for research and teaching model is aimed at simultaneously increas- ery (aerial photography and topographic applications. We describe the present state of ing the effi ciency of fi eld work while adding shaded reliefs) and digital elevation models, fi eld mapping technology from our experience, the attributing capability of GIS to develop fi eld work should evolve into an iterative and discuss the advantages and disadvantages fi eld data products that are more data rich process where linework is roughed out of different approaches and technologies. Key than any paper map could ever achieve. We in the fi eld, refi ned during evening fi eld ses- to this discussion is a philosophical difference emphasize three basic rules in the develop- sions, then potentially revisited if problems in attitudes toward how data are collected and ment of this data structure. (1) A fi eld GIS arise. This procedure is particularly effi - organized. Finally, we provide some thoughts map should emphasize line and point objects, cient when three-dimensional visualization is on how we might proceed in the future. avoiding polygons, objects that can easily be added to the system, a feature that will soon One important lesson we have learned is that constructed outside of the fi eld environment. become the norm rather than the exception. no single system is perfect for all applications, (2) Keep it simple stupid (KISS) is a critical We note that using these systems is particu- and the worst systems are ones that seem logical rule for setting up data structures to avoid larly important for future developments in in the laboratory but are totally impractical in fi eld GIS systems that are less effi cient than metamorphic geology, sedimentary geology, a fi eld situation, or vice versa. We discuss how paper. (3) Data structures need to develop a and astrogeology, but other applications are these technologies can be applied to metamor- compromise between display and data entry, clearly also possible. For geoscience instruc- phic geology as well as economic geology, and with display always trumping data entry tors who teach fi eld geology classes, we note describe how further development of these tech- because geologic insight is the primary goal. that it is critical that these systems be incor- nologies will be crucial to future planetary geol- This paper contains two sample blank data- porated into all geoscience fi eld programs, ogy expeditions. We then consider the similarly bases that illustrate these approaches for two but research is needed on the best teaching revolutionary impact the technology can have applications: (1) generic bedrock geologic approaches in the use of the technology. on teaching fi eld geology. mapping, and (2) metamorphic geology map- ping multiple generations of fabrics. Key fea- INTRODUCTION HISTORY OF THE TECHNOLOGY tures in our approach are to use display as a fi rst-order attribute, sorting point objects Until recently, the tools of the fi eld Until the mid-1990s, computer systems were into four basic types (station, orientation, have seen little change since the nineteenth cen- not practical in the fi eld for any application sample, photo) and lines into the four basic tury: a hammer, a hand lens, and a geologic com- other than simple data logging or for support of contact types (depositional contact, uncon- pass with inclinometer. We have now entered geophysical surveys, the latter typically associ- formity, intrusive contact, ), plus other a new era in technology where rugged, light- ated with substantial logistical support (Table specialized data layers where needed. Indi- weight fi eld computers, geographic information 1). The problems with early portable computer vidual GIS objects are further attributed, but system (GIS) software, global positioning sys- systems were fourfold.

Geosphere; June 2010; v. 6; no. 3; p. 275–294; doi: 10.1130/GES00503.1; 7 fi gures; 2 tables; 2 supplemental fi les.

For permission to copy, contact [email protected] 275 © 2010 Geological Society of America

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

1. Power. The battery life of laptops of the resistant as is common with modern, rugge- With the introduction of the fi rst handheld time was far too short for a full day in the fi eld dized fi eld computers. computers in the early 1990s, many of these without large packs of spare batteries. More- 3. Displays. All of the early liquid crystal dis- issues were overcome with a device of man- over, the lead-acid and nickel-cadmium bat- plays (LCD) became unreadable in bright sunlight. ageable size and heft, a long battery life, and teries in common use at the time were heavy, 4. Processing and communication. Process- a satisfactory (initially monochrome), outdoor- bulky, and slow to recharge. ing power for early portable computers was readable display. That development led many to 2. Portability. Computers were too heavy insuffi cient to run even the modest GIS software consider using these devices by the mid-1990s and bulky to carry the large distances geolo- that was available. In addition, wireless commu- (Table 2). gists travel on foot. Furthermore, the hardware nication and GPS positioning technologies were With the present generation of devices was fragile and not impact, water, and dust not widely available. and software, computerized fi eld mapping is

TABLE 1. HISTORY OF DEVELOPMENT OF MODERN FIELD SYSTEMS Year Hardware Software development Experiments by authors* Experiments by others development First reasonably priced GPS Software developers work in GIS used by many for offi ce 1990–1991 GIS and computer graphic None devices become map development available fi elds ESRI introduces Arcview GIS; MapInfo Introduction of fi rst and Intergraph GIS Limited use of GPS GIS becomes widespread 1992 pen computer alternatives; handwriting- handhelds and GIS as mapping tool, but no (GoPoint) recognition software software fi eld usage introduced with pen computers 1993 First PDA (Apple Continued developments No new experiences GIS usage continues Newton) in GIS Geological Survey of Pavlis writes “I have a Early pen-based Canada develops dream” essay on potential Geological Survey of 1994–1995 devices rejected in Fieldlog system; Penmap of technology for fi eld Canada begins fi eld mass market development for pen applications with Fieldlog computers applications Palm Pilot and Windows CE Initial testing of software and Initial experiments by 1996 PDAs introduced; Penmap initial developments hardware several groups Win95 pen tablets available Transfl ective J.D. Walker (UK) and B. Brimhall (UC Berkley) (sunlight) GIS software improves; First fi eld experiments with screens appear; ArcGIS 8 (1999) and begin fi eld experiments handheld devices using 1997–2000 pen computers ArcPad 5 (2000) Fieldlog and Penmap in with Arcview and Penmap, improve; digital introduced; GPS respectively; many groups cameras become incorporated into GIS Alaskan and Death Valley experiment with various affordable for software fi eld work dataloggers (e.g., Palm, routine fi eld work WinCE, Psion) Abandoned AutoCad-based UC Berkeley group ArcPad 6 together with Fieldlog applications in continues development Tablet PC for favor of ArcPad-ArcGIS of Geomapper; UK group Wintel (1992); ArcGIS 9 become a viable fi eld tool addressing major system; extensive develops new tools for fi rst true outdoor limitations in ArcPad 5; hardware and software ArcGIS 8/9; development 2001–2005 screens become laser ranging devices experiments through of MapIT for geologic widely available; become available for Keck Foundation funding; mapping using TabletPC; ultramobile (<1.5 targetless surveying of interrupted by Hurricane Geopad group formed kg) PCs appear steep terrain Katrina; began paperless and develops working fi eld work in 2004 fi eld system Microsoft Origami ArcPad 7 solves several Experiments with numerous project develops issues in ArcGIS-based data schemes for ultramobile PCs systems, but introduces different types of fi eld Continued development 2006– (future of fi eld other issues; three- operations; new equipment of the Geopad project; present technology); dimensional visualization experimentation through extensive work in Europe smartphones tools become more UT system grant in 2006; using various systems begin replacing widespread with potential two UTEP fi eld classes PDAs for fi eld applications taught virtually paperless Full three-dimensional Ultramobile PCs replace PDAs; data acquisition and Predictions electronic visualization will be Plan to abandon use of for future make incorporated into fi eld Brunton compass within Brunton compass systems; traditional maps, two years like compass, will be a museum piece limited to library archives Note: PC—personal computer; PDA—personal digital assistant; GIS—geographic information system; GPS—global positioning system; UTEP—University of Texas at El Paso; UC—University of California; UK—University of Kansas. *Terry L. Pavlis, Richard Langford, Jose Hurtado, and Laura Serpa.

276 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

straightforward and routine. A critical devel- southern Alaska (e.g., Bruhn et al., 2004; Pav- sion of the mobile GIS package, ArcPad. opment has also been the commoditization of lis et al., 2004). The dusty, hot dry weather of Together with ArcGIS, this software system GPS positioning devices, which now allow Death Valley and the cloudy, cold, rainy weather combined a simplifi ed software interface for spatial precision not possible as recently as in southern Alaska provided a range of condi- fi eld-based data acquisition with a sophisticated the early 1990s. Other recent developments tions in which to test the survivability of the GIS for data manipulation in the laboratory or (Table 2) like automated compass and/or hardware. Because of large power needs of this camp. With the release of ArcPad it was also inclinometers, increasing availability of light hardware and the remote wilderness areas in possible to assemble an affordable system for detection and ranging (LIDAR) data sets, and which we were operating, painful power avail- instruction, and as a result, we and others began high-resolution satellite imagery suggest this ability issues, which we continue to face today, to experiment with teaching using handheld is only the beginning. became apparent. It is interesting that in these devices (e.g., Clegg et al., 2006). At the time of In the mid-1990s, two of us (Pavlis and fi eld projects we never used ruggedized devices, this writing we have collectively used variations Serpa) began experimenting with systems that yet over the course of three fi eld seasons on both of this system teaching fi eld geology classes at combined handhelds for fi eld data acquisition projects we experienced only one device failure, the University of New Orleans, Massachusetts and standard laptops (including early pen tab- amounting to only a one-day data loss. Institute of Technology, and University of Texas lets) for data compilation (Table 2). This early By the late 1990s, a software package called at El Paso (UTEP) with groups ranging between system used Fieldlog software developed at the Penmap (www.Penmap.com) became available. 7 and 15 students. We have also used the sys- Geological Survey of Canada along with a com- We attempted to use this software as well as a tem in day-long fi eld exercises as part of several mercial software package, Fieldworker, for the later customized geologic graphical user inter- other classes in groups as large as 22 students. Apple Newton handheld (Brodaric, 2004; Bro- face (GUI) called Geomapper (Brimhall and Most important, however, we have contin- daric et al., 2004). This system emphasized a Vanegas, 2001; Brimhall et al., 2002, 2006). We ued to use the technology in our fi eld research point-based GIS approach where the handhelds also experimented with using Fieldlog on newly activities in dozens of fi eld-based projects, and and Fieldworker were used for recording point- available pen tablets in the fi eld. Although they paperless fi eld work is now the norm, not the based observations, and the graphical tasks of both seemed to be ideal tools, we ultimately exception. We have so thoroughly adapted to the mapping were typically done on paper and later rejected both Fieldlog and Penmap largely system that we now tend to get frustrated with digitized using the Fieldlog application. Field- because of the limitations of pen tablet systems, paper mapping as highly ineffi cient with very log was essentially a series of applets running including (e.g., Pavlis, 1999; Pavlis and Little, limiting capabilities. under a computer-aided design (CAD) program 2001) their high cost and excessive weight, and Collectively, our experience is broad enough for and graphics management. The the continued lack of outdoor-readable displays. that we believe we have probably made most advantage of the system was that the software These problems remain to this day. In addition, of the likely mistakes one can make in using was designed with a simple interface for fi eld Penmap was a relatively “buggy” piece of soft- these systems, and we have developed a practi- applications and was intended to insulate the ware and the Geomapper GUI emphasized a cal approach that is exportable to virtually any user from the complications of a full-blown polygon input system that was inconsistent with academic institution. Some key lessons that we GIS system (e.g., see Brodaric, 2004; Pavlis and the approach we had developed independently have learned from our experimentation include Little, 2001). (see following discussion). the following. During the late 1990s we experimented with From our perspective, an important advance 1. The technology has advanced rapidly in Fieldlog during two research projects, one for fi eld data acquisition came in 2000 when the past 5 yr, and even since Clegg et al. (2006) in Death Valley (e.g., see Guest et al., 2003; ESRI (formerly Environmental Systems evaluated these systems. Thus, if someone was Golding-Luckow et al., 2005) and the other in Research Institute, Inc.) released the fi rst ver- introduced to early versions of the technology

TABLE 2. HISTORY OF TECHNOLOGY LEADING TO MODERN FIELD SYSTEMS Year(s) Computer hardware Technologies applied to fi eld Other relevant technologic Computer applications in geology developments geosciences Early Aerial photography and twentieth Not yet invented derivative topographic maps + Major technology advances in None transportation revolutionize fi eld numerous subfi elds century work Background work in computer First use of computers in 1950s– Primitive computers None geosciences, primarily 1960s graphics geophysics Early computer graphics (e.g., Key developments in First stereonet plots, graphical data); computer graphics lead First extensive computer- 1970s microprocessor computers no direct fi eld applications to GIS; background work based applications begin outside of geophysics leading to GPS Personal computers become First public domain GIS widespread; Computer graphics become application (GRASS) and Early computer revolution early laptops widespread for drafting; fi eld commercial GIS software changes virtually all 1980s appear; Psion use not yet practical because of in 1982—ESRI’s Arcinfo areas of geosciences defi nes PDA limitations in both hardware and and Intergraph applications; with personal computer and introduces software GPS perfected by U.S. applications fi rst handheld Department of Defense computer Note: PDA— personal digital assistant; GIS—geographic information system; GPS—global positioning system.

Geosphere, June 2010 277

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

and found it lacking, we suggest people take a Windows XP tablet personal computers (PCs) experimentation for fi eld applications and in new look at the systems. with ArcPad and ArcGIS for use as both fi eld teaching, we have become comfortable with 2. There are still important, unresolved tools and offi ce and/or camp tools (Fig. 1). using them for research. In the future, this may issues, particularly software limitations, and We also use ArcGIS and ArcPad Studio (also change (Table 1) as tablet computers become anyone interested in this technology needs to called ArcPad Application builder) software smaller and cost less (e.g. ultramobile PC’s, recognize the time commitments required (see on the tablet PCs to create and export GIS netbooks, and the iPad tablet), and smart- discussion following). project fi les to the handheld units. We rarely phones replace PDA (personal digital assis- 3. Our approach is just one of several, paral- use ArcGIS directly as a fi eld-mapping tool, tant) type handhelds. In many respects the new lel approaches to using this technology that have preferring to use it to edit maps in camp or small tablet devices are an ideal compromise been undertaken. Specifi cally, most of our expe- at the offi ce, and generally use ArcPad as the because they are lightweight (<1.5 kg) and rience has been in an academic context and we primary fi eld tool. This approach is similar small enough to carry easily, yet their screens have emphasized systems that are sustainable to Geopad (http://geopad.org/) and projects are large enough (7–15 cm) to display more without a large fi nancial commitment. described by Clegg et al. (2006), but our effort information than a handheld. In addition, the is an independently developed system empha- processing power and storage of these devices WORKING FIELD DATA sizing handhelds over pen tablets. is now suffi cient enough that higher end visu- COLLECTION MODEL For research applications, most of us have alization software can run on these machines, also chosen the handheld over a tablet PC. The and companies like Midland Valley are port- Hardware and Software main reason for this is convenience. Compared ing a version of their Move software for fi eld to tablet PCs, small, lightweight handhelds applications because of these developments We use a system that employs Windows are much more portable. In addition, because (see www.Midlandvalley.com). Conversely, Mobile handheld devices running ArcPad and we used handhelds from the beginning of our the transition from PDAs to smartphones is a

“squid” for charging multiple devices

convertible mobile keyboards Tablet PCs ultramobile

Recording Inclinometer c ompass-

12 inverter tablet

rugged units

laser binoculars CF units

handheld bluetooth units submeter Field Cases computers/PDAs unit GPS units

Figure 1. Photograph showing the present generation of equipment we are using at the University of Texas at El Paso. Note that this is a sample of the type of equipment available today, and represents a range of general types of equipment needed for a fi eld effort. GPS—global positioning system; PC—personal computer; PDA—personal digital assistant.

278 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

negative development because of the smaller Data Structure and Field Workfl ow small increments of lost time can prolong a fi eld screen size of most smart phones and the popu- project, or lead to poorer results because critical larity of stylus free touch screens on many of An important characteristic of any GIS is that areas might not be visited. As a consequence, these devices. data (attributes or linked information) can be we believe the effi ciency of the data scheme For software, we have become fi rm believ- attached to features or groups of features on a used in fi eld data collection systems is critical to ers in not using a fully featured GIS program map. The ability to attach information to map the success of fi eld GIS mapping. In particular, like ArcGIS as a fi eld tool. Although ArcGIS objects gives the user the ability to change the three major issues are apparent.1 is very versatile, its versatility can be a curse display based on attributes, overlay transpar- 1. Polygon objects should be avoided. This in the fi eld because features like cascading ent layers, or even visualize the data in three is because they are too dynamic to build in the menus, right-click requirements, and complex dimensions (3D), all of which represents a huge fi eld and may require constant reconstruction options can produce confusion for anyone but advance in our ability to solve fi eld problems. as mapping progresses. There are a few cases experienced users of the software. Customized Although these features are the real strength where this rule might be eased, e.g., in surfi cial interfaces like that employed in the Geopad of GIS, geologists often treat GIS as simply a mapping or bedrock mapping with scattered project (http://geopad.org/) and that of Walker method for drawing maps. To this day, most dig- outcrops. Our experience, however, is that using and Black (2000) and Black and Walker (2001) ital releases do not take advantage polygon objects to build a map always leads to can improve ArcGIS as a fi eld tool, but in our of the true power of GIS because they typically diffi culties because of unforeseen topologies opinion, the program generally is too user hos- use minimal attributing. This is particularly that invariably develop. Thus, we use labeling tile for the fi eld environment. From our per- true of most regional maps digitized from older or even hand-coloring of printouts during fi eld spective, ArcPad represents a better fi eld tool paper maps where little, if any, useful informa- work and only build polygons as derivatives because it can be easily customized for the fi eld tion is attached to the map that was not available after the fi eld work is completed. environment, minimizing the potential for an on the paper map. A very different type of map 2. Keep it simple stupid, or KISS, is an inexperienced user to become overwhelmed by can be built, however, when the data are col- important rule for fi eld GIS systems. Essen- the software. However, ArcPad is still fl exible lected in digital form in the fi eld with attributing tially, this means simplifying all fi eld data entry enough that it can be adapted in real time in the capabilities of GIS in mind. to the minimum amount needed for the project. fi eld. Moreover, while most people can learn The fi eld GIS approach introduces a wide Our rule of thumb is that point objects (e.g., sta- the basics of ArcPad in a single 2–3 h training variation in possible procedures for fi eld work. tions) should contain no more than 6–8 attribute session, few new users would feel even mar- Setting up a data structure that takes advantage fi elds for data entry, and most of these should ginally comfortable with ArcGIS in that time of the GIS framework is nontrivial, and will not be fi elds that are required every time a data frame. Nevertheless, it is important to realize require new fi eld habits (e.g., Brodaric et al., point is collected. Examples of high-priority that with ArcPad it is essential that at least one 2004). The challenge to a fi eld geologist is to attributes in a station fi le might be: station iden- person in a fi eld party has the technical ability take advantage of this capability while not get- tifi cation (ID), geologist, date, location method, to build a project in ArcGIS and ArcPad Studio, ting bogged down in collecting too much data or and a note. For line objects an even stricter rule to export the project to the handhelds, and to ending up with an unworkable mass of poorly should be enforced; line objects should contain debug occasional software and hardware prob- organized data (Asch, 2003, 2005). no more than 3–4 attributes. This rule for line lems. However, this is generally signifi cantly Early experience with the Fieldlog system objects arises because attributing can be dis- easier to achieve than assembling an entire fi eld quickly taught us that a data structure set up tracting when drawing the linework on a map. party that is familiar with ArcGIS. for a laboratory environment is not necessarily For example, consider the process of drawing An alternative software product, Map IT workable in a fi eld environment. Specifi cally, a a contact that is intermittently exposed with (http://www.uniurb.it/ISDA/MAPIT/) was data manager in the offi ce might imagine a data several covered intervals. On paper, this can be developed in Europe, and Clegg et al. (2006) collection scheme that includes a standardized performed with little thought. However, in GIS evaluated its utility for various applications set of data that can be rapidly sorted and ana- the simplest method for graphic approach is to compared to using ArcPad of ca. 2005 vintage. lyzed in the offi ce. Unfortunately, if that data assign a “quality” attribute fi eld (see sample We have not experimented with the software, in scheme is too complex, even the most patient data), and when a new set of attributes must large part because it requires use of a tablet PC, fi eld person will quickly abandon it if each stop be entered for each new line segment, it can be but the program has been used in the Geopad involves data entry into multiple pull-down frustrating. A second rule is that any pull-down project (http://geopad.org/). From descriptions, menus with complex pick lists. Thus, all data menus with pick lists must contain no more than it appears to be a better choice for geologic schemes are ultimately a compromise between 6–10 options with no cascading (nested) pull- mapping applications than ArcGIS, because the critical data to be attached to map objects down menus. Finally, to overcome any limita- like ArcPad, it has a simplifi ed GIS front end. in the fi eld and the other data that can be later tions these guidelines impose, the simple inclu- Clegg et al. (2006) considered Map IT supe- entered in the laboratory, saving valuable fi eld sion of a “note” fi eld, with adequate space for rior to ArcPad, and although we consider some time. This last point is critical, and often misun- generic data entry not covered in the fi xed data of the same issues here, their comparison is derstood by geologists who have done little fi eld fi elds, often proves invaluable. already somewhat outdated in the latest ver- work. Field time is extremely valuable, not only 3. Set up the data structure to allow a natu- sions of ArcPad. because of real monetary costs, but also because ral workfl ow in the fi eld. It is crucial to develop

1Supplemental Files 1 and 2. Zipped fi les containing examples of two generic shapefi les that use the data structure described in this paper. The fi les contain two generic projects that, when unpacked, can be used directly in ArcPad 7 or 8. See Appendix 2 for details of working with these fi les. If you are viewing the PDF of this paper or reading it offl ine, please visit the full-text article on www.gsapubs.org or http://dx.doi.org/10.1130/GES00503.S1 and http://dx.doi.org/10.1130/GES00503 .S2 to view Supplemental File 1 and Supplemental File 2.

Geosphere, June 2010 279

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

a compromise between display and data entry Our typical “orientation” data object (Figs. a station code, position, date, the person mak- issues, display always trumping data entry. GIS 2 and 3; Appendix 2) illustrates one of these ing the observation. Other data are then linked systems allow data sorting by layers and by principles. In this fi le we limit data entry to a to this data layer by a common station code, attributing data, i.e., assigning values to a data single with the option of including which is written to all associated data tables. In table. Data layers are a fi rst-order attribute and a trend and plunge measurement, and include an many ways, this is the ideally fl exible system represent one simple way to separate informa- attribute fi eld to record the type of orientation and emphasizes the capability of modern data- tion, but excessive use of data layers can also be data being measured (e.g., bedding versus cleav- base systems (Brodaric, 2004). This approach confusing. For most geologic operations we use age) and a note fi eld to record any subsidiary was used in both a successor to Fieldlog (Geo- four types of point objects and 4–10 line types information. The orientation data are then dis- fi eld) developed at the Yukon Geological Sur- as data layers (Fig. 2). The point objects are sta- played directly on the map using layer defi ni- vey (Lipovsky et al., 2003) and applets devel- tions for noting observations, sample locations, tions imported from ArcGIS. In areas of rela- oped for ArcPad 6 (Geologic Data Assistant, structural measurements, and photographs. The tively simple structure with only sedimentary GDA) developed at the U.S. Geological Survey line objects always include the four basic types bedding, this layer also can serve as the primary (Thoms and Haugerud, 2006). Nonetheless, of geologic contacts: depositional, unconfor- data entry form, e.g., a quick record of orienta- Murphy’s law (i.e., the adage, “anything that mity, intrusive, and faults. In addition, other tion data to be plotted on the map with a short can go wrong will go wrong”) always applies structural elements, e.g., traces, thin note on rock unit, or other information. to fi eld projects, and we have become hesitant dikes, and axial traces, can be added as lay- In more complex areas, however, this simple to use this approach because of a hidden poten- ers when needed. These point and line elements format may be insuffi cient. For example, in tial for data loss. Specifi cally, because all other are the natural building blocks of geologic metamorphosed rocks, multiple orientation mea- point data have their spatial reference linked mapping, and we typically include each type surements might be made in a small area, and to the station fi le, if that single fi le were to be as a separate data layer, and each data layer is although these data are needed, plotting all these corrupted there is a high potential for complete then attributed further (Fig. 2). A sample blank data on the map soon produces confusion. With data loss. Routine backups minimize this issue, geodatabase and set of shape fi les using this data the recent development of recording digital com- but when using this approach in both GDA and structure are included herein as an appendix. pass and/or inclinometers (http://www.gsinet Geofi eld we had partial data losses that required .co.jp/english/geoclino/index.html), the volume hours of work to reconstruct. Point Data of orientation data obtainable in these types of Because of these issues, our recommendation For many studies virtually all fi eld data are studies can rapidly grow to unmanageable size. is that station point objects should be set up with point observations. Even in areas of exten- In such cases, we choose which measurement to a specifi c project in mind, with at least one point sive outcrop, point-based observations are an display using the orientation layer, and we rel- object fi le aimed at the specifi cs of the proj- important part of the fi eld data set, and signifi - egate all the other data to either a separate data ect. Thus, for projects emphasizing sampling, cant fi eld time has traditionally been devoted fi le that is later linked into the database, or to there should be a specifi c sample fi le, and the to point-based observations. This is the tradi- the station table. Alternatively, a Boolean “plot other data can be relegated to one or two other tional station approach, in which fi eld notes to map” fi eld (Fig. 3) can be incorporated into objects. Similarly, for bedrock mapping, orien- are keyed to a specifi c map position. We have the data structure in ArcGIS. Unfortunately, tation information is often a focus of the work, experimented more with point-based GIS than plotting in the present version of ArcPad is and thus we typically separate that object in this any other part of the data structure, and our phi- restricted to single attributes, and thus this type type of work. Notes tagged to these fi les can losophy is infl uenced strongly by the approach of more complex ArcGIS layer defi nition cannot also take different forms: text fi elds as a fi eld taken by Brodaric (2004). Nonetheless, formu- be directly exported to ArcPad. in the attribute table; linked data fi les defi ned as lating rules for this part of the data structure Other point-based data produce different attributes with the linked fi les representing text puzzle remains diffi cult. This arises in large part challenges. Foremost among these is the tradi- fi les, photos, sketches, hand-written notes; and because of different styles among fi eld workers; tional station, which can serve many purposes, separate data tables. thus, it is diffi cult to establish a data structure depending on the fi eld project and the individual that works well for everyone. The data structure (Brodaric, 2004). Most geologists are familiar Line Data shown in Figures 2 and 3 is developed primarily with this type of data object from paper map- Most linework is given a primary attribute for bedrock mapping, but we believe the basic ping, but input tends to be a freeform personal for outcrop quality (exposed, approximate, or form of the data structure can be easily molded preference. What constitutes a station varies concealed). We use that attribute as a primary to other applications. Specifi cally, although we among individuals. For example, many fi eld display parameter in part because of a long tra- recognize the common use of four types of point geologists number all stops as stations, while dition in geology. However, there is no particu- data (station notes, photos, samples, and orien- others may use a different code for sample lar reason that this exact approach be used other tations), we typically combine the fi rst three into locations versus notes versus orientation mea- than tradition, and the fi eld GIS can allow a more a single station layer and store orientations in a surements versus photographs, whereas others fl exible approach. We have used an expanded separate layer so that they can be more easily may mix all four in various permutations. Thus, list, containing exposed, exposed_projected, plotted as strike and dip symbols (Fig. 3). From although we typically combine station notes, approx_fl oat, approx_fl oat_projected, and our experience, most of the other point data is photo descriptions, and samples into one point inferred, the “projected” implying a contact that archival information that is accessed later and is fi le, this is largely a personal preference. One is extended using aerial photography or sketch- not useful as part of the fi eld map display. Dur- approach to this issue is the approach used by ing from a distance. We have experimented with ing routine mapping we often make the station Brodaric (2004) in Fieldlog. A station object a quantitative attribute for contact quality (e.g., layer invisible, the computer mapping equiva- in this approach is the place-keeper for spa- <10 m, 30 m, inferred), but found that this pro- lent of using pinholes in paper maps with station tial position, but the data table for the object is cedure confused people familiar with the con- numbers written on the back of the map. largely composed of metadata, including, e.g., ventional map scheme.

280 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

Point Shapefiles

Line Shapefiles

Figure 2. Examples of typical data entry screens for our ArcPad projects. See text for discussion.

Geosphere, June 2010 281

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al. Generic Station Table Data Structure

Station feature class (spatial position typically captured by GPS)

p. 3-5: Notes p.1: principal facts p. 2: linked information (250 character fields) Stationnumber___ 4photofilename ____ 1Date___ photo caption____ 1 Geologist___ sample______2LocationMethod___ sample desc____ 2 Outcropquality___ 3long note (y/n) other options: 4longnotefile____ map unit, weather, note1 etc. note2 note3

1 3 Can be automated Boolean (y or n field) Limitations: with this data structure multiple 4 samples or photos at a station need to be 2 Pull down menu item File name generally not completed in field recorded in notes, with information linked later in the lab. Note: longnote file name refers to the file name of notes entered in word processor +/- sketches embedded in the long note. File name can be entered, but commonly will need to be linked to the database later, not in the field Figure 3. Illustration of alter- native data structures for Alternative 1: Sample or photo intensive stations station fi les. Example fi les included with this paper use p. 3-5: Notes the data structure illustrated p.1: principal facts p. 2: linked information (250 character fields) at the top of the fi gure, but the Stationnumber___ # of photos___ p. 3 photo descriptions other alternatives have been 1Date___ # of samples_____ 1 used successfully. GPS—global Geologist___ 3 2 LocationMethod___ long note (y/n) p. 4 sample descriptions positioning system. 2 Outcropquality___ 4 longnotefile____ other options: map unit, weather, etc. p. 5 short note

1 3 Can be automated Boolean (y or n field) 4 2 Pull down menu item File name generally not completed in field

Alternative 2: Orientation data intensive stations (e.g. detailed structural studies)

p. 3-5: Notes p.1: principal facts p. 2: linked information (250 character fields)

Stationnumber___ # of photos___ p. 3 photo descriptions 1 Date___ # of samples_____ 1 Geologist___ 3 long note (y/n) 2 LocationMethod___ 4 longnotefile____ p. 4 sample descriptions 2 Outcropquality___ 4,5 other options: orientation data map unit, weather, table ____ etc. p. 5 short note

1 3 5 Can be automated Boolean (y or n field) Linked orientation data could be 4 2 Pull down menu item File name generally not generated as a spreadsheet, text, or completed in field database table from another application; Can also be ignored and “longnote” used in place of a separate data table, depending on the application needed

282 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

Depending on the project being undertaken, map, implementing a fi eld GIS system for meta- and the project’s specifi c needs, in determining a number of other attributes can be given to morphic and economic geology work strongly how much effort is spent on daily data compi- linework. For mapping in sedimentary rocks underscores the KISS rule. We have typically lations and/or editing versus fi eld data entry (e.g., Figs. 2 and 3) we have commonly used settled on a simplifi ed data structure with a data versus post–fi eld work data compilation and/or a “contact type” attribute. In an area with well- layer for each linetype (e.g., S0, S1, S2), each editing. If a fi eld project is working in a remote established stratigraphic units, this attribute can with no more than two basic attributes (quality site, evening fi eld time may be limited, and thus, be used to defi ne the standard contact, but for and note), and with no requirement that these required evening data compilation and/or editing areas with poor stratigraphic control or thick attributes be set in the fi eld. In practice, to speed should be minimized. Alternatively, if a project units, we also include a “sedimentary_ bedding_ fi eld operations, we generally do not attach any has extensive logistical support (offi ce and/or trace” attribute to identify the line as an intra- attributes for foliation traces other than layering room for evening work with no power limita- formational bedding surface trace distinct from and we use notes or photographs to clarify the tions), fi eld efforts can be maximized by delay- a traditional formational boundary (Fig. 4). For quality of the mapping. In this case, attributes ing many steps to evening data compilation. detailed sedimentary studies, sequence bound- can be added later, including linework groupings Similarly, there are hybrid fi eld procedures aries, fl ooding surfaces, and hierarchical bound- in an ArcGIS geodatabase, to clarify the nature that may be appropriate, depending on indi- ing surfaces can also be added. The ability to of the information. For fold systems it is tempt- vidual preferences. For example, we have found zoom into a small area is particularly important ing to develop a complex data structure that many people prefer to use a traditional paper for mapping smaller scale features, such as would encompass the entire range of possible notebook in the fi eld for developing sketches bounding surfaces, because until now, these sur- fold types, a strategy that in practice becomes and notes, minimizing data entry in the fi eld. faces typically have been traced only on outcrop unworkable. Thus, for folds we use only one For teaching applications this may be a pre- photos. Alternatively, in volcanic rocks, or areas layer for fold axial traces, typically limiting the ferred technique to encourage good note taking. with volcanic rocks interbedded with sedimen- fi elds to an attribute for graphical display. In However, when this type of hybrid approach tary rocks, the contact type attribute can be used our sample data fi le, for example, we use only is used, it magnifi es the need for evening data to distinguish volcanic versus sedimentary con- a fold form attribute, and other data (including compilations; either forcing evening data entry tacts. We always include a note attribute with a fold generation, orientations) are entered in a or at least correlation of notebook entries to the large data fi eld where additional information can long note fi eld. An alternative data structure is digital data fi les. be added. This information is the type of narra- to use both a form and generation attribute, but tive or descriptive material a geologist would the assignment of folds to generations routinely Logistical Issues for Computer-Based traditionally write in their fi eld notebook. It is can cause confusion if this approach is used in Field Projects interesting that having this recorded digitally as a new fi eld area where generation assignments an attribute of a specifi c GIS object ultimately may be in fl ux for days. Field geology projects are often undertaken changes many people’s note-taking habits. Spe- in remote areas where electrical power is limited cifi cally, the note fi eld for an individual line Data Compilation and Note Correlation (or nonexistent) and where environmental con- element can be used to add information about ditions can take a harsh toll on electronics. Both the feature that could never be done with paper An important element in GIS-based fi eld of these issues are closely related. mapping. For example, notes like “this con- studies is compilation at the end of the day, Since most hardware has, at most, three tact is speculative” or “this is accurate to ~1 m which includes both routine backups and data days of useful battery life, a fi eld party needs through GPS positioning and high-resolution entry and/or repair. The procedure can be as to develop a plan for keeping equipment opera- aerial photography,” are possible. simple as downloading a digital camera, chang- tional that includes redundancies in the event Working in metamorphic rocks leads to a ing fi le names, and linking the photos to the of equipment failures. In group exercises for somewhat different linework data structure (Fig. database. However, more complex procedures geology fi eld classes, this may require some- 5), where more complex attributing may be can be developed, including revision of line- thing as elaborate as a generator, and multiple needed to allow a particular type of data display. work using aerial photography, digital eleva- power outlets for charging devices. For smaller Similar issues also arise in economic geology tion model (DEM) shaded reliefs, or both, as groups or applications in extremely remote (Brimhall et al., 2006). For example, in com- well as editing attribute tables. These editing areas, the best power solution we have found is plexly deformed metamorphic , in addi- operations are all GIS-related functions that solar-powered, portable 12 V charging systems. tion to primary compositional layering, there is can be done outside the fi eld environment, and Careful consideration, however, must be placed typically more than one structural fabric (e.g., the extent of this exercise depends on the proj- to balance space and/or weight limitations with S0, S1, S2) and surface traces of these features ect, the personnel in a project, and the preferred power needs. This is problematic since fi rst- can be routinely mapped. It may also be desir- procedures for the group. This editorial step, time users are unfamiliar with how to budget the able to map other structural elements (e.g., axial however, is critical, and is the digital equivalent wattage per day required, the critical criterion traces of multiple fold generations, linear fabric of “inking of lines” or daily map compilations. for selecting the size solar panel required. Simi- traces), and features such as metamorphic iso- Conceptually, this step can be very important larly, most car batteries are a poor choice for a grads or hydrothermal alteration zones. Because for both students and a researcher as a refl ec- solar charging system because they are gener- these fi eld GIS systems allow us to simultane- tion on the day’s work, is crucial in planning ally prohibited in aviation cargo, cannot tolerate ously map and selectively display these multiple for the next day, and can be a critical meta- deep-cycling, and they typically have a much surfaces, it is in these environments where we cognitive step in fi eld problem solving (e.g., larger capacity than needed. From our experi- have found the most profound improvement http://serc.carleton.edu/NAGTWorkshops/ ence, a fi eld party of 2–4 can be maintained with over traditional paper-based mapping. metacognition/index.html). a sealed 10–15Ahr battery charged with a 20– Because there are so many lines (each poten- In our experience, it is important to evalu- 40 W solar panel. A critical piece of electronics tially with several attributes) being drawn on a ate the fi eld project at hand, the fi eld personnel, that is often overlooked in constructing a solar

Geosphere, June 2010 283

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al. 53

48

45

42

. 42 86 33 ( 37 46 o . 21 69 79 89 89 .

62

59

85

40 80 73 87 87

Meters 1:12,000 300150 0 300

Figure 4. Example of an ArcGIS map generated using our techniques. Figure illustrates the power of digital mapping to develop extensive linework on a map beyond traditional mapping, here illustrated as a bedding-plane-trace map. Note that in this area, mapping of only the traditional forma- tion boundaries (heavy black lines) and faults would have produced a map with little useful information, whereas inclusion of the bedding traces shows the structure in detail. The map is from southern Alaska in the fold-thrust belt of the St. Elias Range, just east of Bering Glacier (our data).

284 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

charging system is a voltage regulator to avoid spent several days of actual mapping in an area “fuzzy data” wherein facts are mixed with inter- destroying a battery by overcharging, a device where they have access to high-tech support— pretation. Second, collecting basic fi eld data is readily obtained from businesses that sell solar software, hardware, or both. The list of potential laborious and, in some cases, dangerous. As a power equipment. problems is long when users are unfamiliar with result, the intellectual property encapsulated by The physical environment (e.g., weather, alti- the system. Thus, new users need to gain experi- a geologic map has a very high value to the pro- tude, wildlife) requires logistical planning simi- ence with the equipment in an area where there ducer of the map, yet to the broader community lar to that required for geophysical fi eld experi- are no time pressures and high-tech resources the information is merely a part of a broader col- ments. Weather issues, for example, affect the are available, not in a remote site with limited or lection of knowledge. kind of power system that is most practical for no communication. We suggest that if all fi eld geologists began a project. For example, preventing short circuits to use the type of GIS system described here in a wet environment is a fundamentally differ- DISCUSSION and elsewhere (e.g., Clegg et al., 2006; Brimhall ent challenge than working in a desert where et al., 2006), the community would ultimately dust and sand might clog terminals and cooling Field geologists have a poor record of devel- develop a different attitude about fi eld data. The fans. Equally important is that someone in the oping collaborative fi eld efforts, a tradition that “fuzzy data” issue can be entirely eliminated fi eld party takes the responsibility of technical goes back to Roderick Murchison and Adam because in a GIS there can be an explicit segre- expert. That person must be knowledgeable Sedgwick’s well-known feud in the nineteenth gation between objective, quantitative data col- enough to maintain all equipment in working century. This problem arises from the basic lection and subjective data interpretation. The order. Appendix 3 gives a simple checklist of nature of traditional fi eld work. First, geologic basic data used to develop a map, such as fi eld useful spare equipment, tools, and suggestions maps are a derivative of a series of point obser- descriptions, photographs, and orientation data, for troubleshooting. vations, each of which can be subject to random can always be extracted from the database. Fur- As a fi nal logistical note, geologists should chance and subjective interpretation (e.g., Ernst, thermore, if the basic fi eld data are always col- not proceed to a remote location until they have 2006). As a result, a geologic map becomes lected in digital form within a GIS, the data are

A

Figure 5 (continued on following page). Example of the power of digital mapping in analysis of the complex structure of metamorphic terranes. (A) ArcPad dialogue boxes using the sample data fi les accompanying this paper.

Geosphere, June 2010 285

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

inherently archival, and ultimately can be easily during which fi eld data should be the sole intel- NCGMP09/) that minimizes required attributes shared with the broader community. This data- lectual property of the scientist who produced in the GIS but allows more extensive attri- base would also allow other workers to directly it. After a project is completed and main results butes in nonstandard fi elds. Although this “one examine the basic observations that support published, however, it is a waste for that infor- size fi ts all” approach is an important step for the geologic interpretations on the map. This is mation to disappear into a fi le drawer, which is regional map compilations in a large organiza- a remarkable advance in how fi eld geology is the typical case today. Thus, it should become tion (Haugerud et al., 2009; Thoms et al., 2009), done because it greatly facilitates the equivalent a routine procedure to archive basic geologic there are undoubtedly complications in detail of reproducing an experiment in other branches data as we move into digital mapping. The ease as these systems evolve. It is important for the of science. of data archival in a GIS format makes this a broad geoscience community to contribute to How fi eld data are archived is beyond the straightforward process, but the data structure these evolving issues because although the stan- scope of this paper, but we believe that it is of the archival information is not. The U.S. dardization of data structure for data release is imperative that this problem is addressed in Geological Survey has developed one archival important, in fi eld systems a standardized data the near future. Clearly there is a time model (http://ngmdb.usgs.gov/Info/standards/ structure may stifl e creativity.

B

Standard Geologic Map with foliation traces Standard Geologic Map with foliation traces and Quaternary removed for clarity

Foliation traces S2 foliation only Foliation traces S3 foliation only

Figure 5 (continued). (B) A series of screen shots of a digital geologic mapping prepared with different data layers that can be turned on or off to display different structural generations (from Pavlis and Sisson, 2003).

286 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

Suggested Changes in Field Procedures clarifi cation of fi eld relationships, and testing of success of this method has not been universal, hypotheses. In a very mature area, fi eld work because one of our recent fi eld classes performed In using fi eld GIS technology we have con- might be limited to refi ning details, but in less poorly with this approach. In this case, the prob- cluded that a number of long-held paradigms well known areas, the fi eld work would require lem appears to have originated when students for fi eld work need to be reexamined. This more extensive geologic mapping. In either became accustomed to using aerial photography extends not only to the research environment case, however, the preliminary work saves and where rock units were clear on the photos, but but also to the way we teach fi eld geology to enhances the productivity of valuable fi eld time. when confronted with an area where photo inter- undergraduates. The iterative nature of collecting and devel- pretation was diffi cult, they experienced major A paradigm of fi eld geology has been that oping fi eld data in a GIS is also different from problems. We do not have a clear solution to this when you leave the fi eld on a given day, all paper mapping. Paper maps are always refi ned problem, and it represents a geoscience educa- linework should be complete. This paradigm is and modifi ed (with pencil and an eraser) during tion issue that needs to be addressed. so fi rmly entrenched that it is a mantra empha- fi eld work, and many people compile the infor- The process of data compilation and fi nal map sized in virtually all fi eld geology classes at the mation nightly to a base map. In the computer preparation is a critical step in fi eld GIS map- undergraduate level. We suggest that although mapping world, this last step, compilation, is ping that, if ignored, can lead to less geologic it is still critical to impress on students the a necessary and potentially major contributor insight than paper-based mapping. Although the need to complete linework and descriptions in to the effi ciency of fi eld work. At a minimum, old-fashioned processes of compilation, ink- the fi eld while the geology is in sight, overem- the tedious work of daily backups of data and ing, and map coloring are tedious, for most of phasis on this concept can be a handicap when cataloging and renaming of data fi les (i.e., pho- us, this process was an intellectually important using modern technology. Specifi cally, in the tos, orientation data fi les) is essential, but other step in data synthesis. It forces appraisal of map interest of increasing fi eld effi ciency, some steps at this stage can be very informative. For patterns and evaluation of accuracy, particularly fi eld tasks are best accomplished as an itera- example, when good aerial photographs are when linked to analyses such as cross-section tive process of fi eld observations, cleanup, and available it is often more effi cient to not worry construction: that is, a metacognitive step simi- further fi eld observations. about precise details like exact contact place- lar to that described above. In a computer-based In conventional paper-based mapping, the ments while in the fi eld, particularly when time mapping system, the tedium of the compilation mapping process began by referring to any is an issue. Instead, roughing out the basic line- and fi nal map preparation step can be largely previous geologic mapping, and these maps work with some careful georeferencing of criti- eliminated. At the same time, however, the data might or might not be carried into the fi eld for cal points may be all that is necessary, and in the synthesis function of this step can be retained appraisal. In the fi eld, the typical procedure evening the map can be cleaned up and refi ned, and enhanced. Along with conventional cross- would be to map directly onto a topographic particularly projected contacts, using the aerial section construction, we have used two other map, aided by aerial photography if available. photography (Fig. 6). Depending on the quality procedures to aid this step: (1) topology build- In some cases, orthorectifi ed aerial photographs of imagery and abilities of the fi eld geologist, ing and editing, and (2) 3D visualization. (or satellite imagery) could be used with or this procedure can usually be completed in less We noted here that we typically avoid poly- without a topographic map overlay, but gener- than an hour each evening. gons during geologic mapping, largely to avoid ally most geologists carry imagery as sepa- This suggestion will undoubtedly appall wasting fi eld time on a process done more eas- rate stereo pairs. The result of having so many many long-time fi eld geology instructors, but if ily in the lab. However, there is also an intel- disparate paper map products was the “fi eld one has not used these systems it is diffi cult to lectual advantage to delaying this step. In the map shuffl e”; i.e., constantly looking from appreciate how effi cient this procedure can be. GIS approach, this step represents the digital old geologic map to topographic base map to In recent undergraduate fi eld geology classes equivalent of coloring a map. Like coloring a aerial photographs, to the landscape in front of at UTEP we were surprised by the dramatic map, this step forces map appraisal line by line you, back to the topographic map, and so on. improvement in the quality of student maps pro- through cleanup of map topology, and forces a Although anyone who has done extensive fi eld duced with this technique. The lead author had close look at the map to aid visualization. The work becomes accustomed to this procedure, it used this technique for some time in research frequency of this step depends on personal is inherently ineffi cient. In contrast, the ability environments, but the impact on a research set- choice. This step can be done nightly if a project to stack multiple, georeferenced data sets on a ting is less obvious than that on students because is fully integrated into a GIS, or it can even be computer screen, including options to make lay- the high skill level of research personnel primar- done continuously if ArcGIS is used as a fi eld ers transparent, avoids the fi eld map shuffl e, and ily led to fi ne tuning of maps in the compilation tool (e.g., Walker and Black, 2000; Black and allows the fi eldworker to concentrate on the task stage. With the lower skill levels of students, Walker, 2001). However, it can also be delayed at hand: understanding the local geologic rela- however, the impact was dramatic because stu- indefi nitely, and if a group is not comfortable tionships. Moreover, when the data are acquired dents quickly saw their errors on high-quality with GIS software, old-fashioned hand-coloring as fi rst-generation digital maps, many compila- imagery when they had time to view their work of printouts can serve the same purpose. tion errors can be directly avoided (e.g., map without the multitasking pressures of the fi eld Finally, 3D visualization systems represent compilation blunders and confl icts discussed by day. We suspect that this improvement is due to the ultimate future for fi eld geology, but at the Campbell et al., 2005). students taking a more active role in planning time of this writing there is no practical sys- The availability of multiple, georeferenced and daily refl ection on their work, an impor- tem that is useful for the fi eld environment. In data layers in a fi eld GIS also leads to a differ- tant metacognitive step in learning (e.g., see particular, true 3D displays are limited to the ent workfl ow. In particular, even before going to http://serc.carleton.edu/NAGTWorkshops/ laboratory environment and the only 3D tools the fi eld, data compilation from existing maps metacognition/index.html), and is consistent available for fi eld work are pseudo-3D applica- and photointerpretation of imagery can lead to with Riggs et al. (2009) observations of student tions using perspective views; e.g., web-based a partially completed geologic map. Field work success when fi eld planning was used by suc- systems like Google Earth, and viewers like can then concentrate on specifi c problem areas, cessful students in fi eld classes. Nonetheless, the iView3D and ArcScene. Similarly, although

Geosphere, June 2010 287

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

Figure 6. Example of raw fi eld maps generated using ArcPad as fi eld tool with evening edits in Arc- GIS. The fi gures show linework plotted on an orthophoto base. A) Linework after completion of a day’s mapping, with pink lines showing bedding traces at the completion of the day’s fi eld work and green lines showing changes in that linework completed that eve- ning (bright red lines—faults; blue lines—unconformities; and map- A ping faults and unconformities shown had been completed during previous work in the area). B) The fi nal map after the evening’s edits. Note how initial fi eld linework was refi ned, particularly in areas of low dip in left-center and lower- central parts of the map. Comple- tion of this linework in the evening saved valuable fi eld time in this area of good exposure, illustrating the power of the technique for fi eld effi ciency as well as accuracy. The map area is in the Indio Moun- tains south of Van Horn, Texas, with an extensional half- in the central part of the map bound by a low-angle, southwest-dipping normal fault (red line) to the northeast and an unconformity to the southwest. The structure below the unconformity is a Meso- zoic fold-thrust system.

B

288 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

some software, like Move from Midland Valley 100 plane-line pairs in less than an hour. This be accomplished during the EVA. In addition to Exploration software (http://www.mve.com/), suggests that unprecedented geometric resolu- the added effi ciency afforded, some tasks that allows true 3D digitizing, this software requires tion can now be done routinely. When com- would otherwise be unfeasible would be made special licensing and is expensive for nonaca- bined with high-resolution GPS this could lead possible. For example, during the Apollo mis- demic units. Nonetheless, Midland Valley is to a whole range of new capabilities in resolv- sions, the astronauts did not have the capabil- developing this software for fi eld applications ing complex geometries. As 3D visualization ity to map geologic contacts and were limited and has signifi cant promise. New software systems develop further, the capabilities of to verbal descriptions of geologic relationships developments in the UK (e.g., Mathers et al., resolving complex structural geometry will and the collection of samples. With a fi eld GIS 2009) and experimentation in the U.S. (e.g., only improve, and we predict a future revolu- system coupled to a geolocation capability anal- Phelps et al., 2009) offer a future for 3D mapping tion in our understanding of complex meta- ogous to GPS, future astronauts could precisely at a variety of scales, but remain an offi ce tool. morphic structures as a result. map out the geometry of geologic relationships Thus, fi eld geology must await future develop- in the fi eld by simply walking them out, or by ments to fully integrate 3D into the fi eld envi- Sedimentary Geology digitizing them from distance with a laser range- ronment, particularly hardware such as heads- The combination of GPS with a fi eld GIS fi nder. Numerical data and descriptions (trans- up displays, 3D displays for mobile devices, also creates new opportunities in sedimentary lated to text via speak recognition), in addition and augmented reality systems. Nonetheless, geology. Sedimentary features are typically to digital photography, could be linked to the 3D visualization can be used as part of evening analyzed as 2D objects. Although the advent of fi eld GIS, just as easily as they can now with map compilation work, depending on the logis- sedimentary architectural studies and sequence ArcPad. In addition, the fi eld GIS system would tics of a fi eld project. For example, we have stratigraphy has highlighted the 3D nature of allow the full use of available remotely sensed used Google Earth to aid in data compilation sedimentary features, understanding of the 3D data as a base map for both scientifi c decision in the evening with classes. In research settings geometry is handicapped by limitations of both making and for navigation. we have used ArcScene and the commercial data collection methods and ability to visualize These enhanced capabilities will also encour- program Fledermaus (http://www.ivs3d.com) the systems. Today, many scientists are experi- age complete documentation and curation of to drape shapefi les and published geologic menting with ground-based LIDAR systems, fi eld data and samples, crucial tasks for activi- maps onto a DEM, and used 3D viewing capa- laser rangefi nder systems, high-precision GPS, ties in locations that are expensive and diffi cult bilities to help recognize mapping errors and or some combination of these technologies. to reach. For example, when fi eld data are col- clean up mapping (Fig. 7). However, these represent only a part of the tools lected in a GIS, (near) real-time transmission needed to best resolve true 3D architecture. and remote analysis and/or archival of that GIS Future Revolution in Field Mapping using a fi eld GIS system allows for data are possible. This can allow the support Geology—Examples huge improvements in the development of true staff on Earth to vicariously explore and better 3D reconstructions. Simple tracing of bedding engage and/or direct the astronauts during any Metamorphic Geology planes becomes a 3D exercise that, in combina- EVA and enable better planning between EVAs. The structural geometry that can be pro- tion with a 3D imaging program, can revolution- This can also allow richer sharing of informa- duced by multiple generations of fold over- ize our understanding of sedimentary systems. tion among astronauts in the fi eld, supplement- printing in metamorphic terranes represents Using these tools together with a GIS mapping ing the voice communications with real-time one of the most diffi cult 3D visualization prob- system represents a powerful combination that data and further increasing effi ciency in the fi eld lems in all of geology. Traditionally one of the is diffi cult to overemphasize. in terms of time, metabolic costs, and amount of main methods for resolving geometry has been ground covered. systematic mapping of the surfaces traces of Planetary Geology different generations of structural fabrics (e.g., Many of the lessons we learn from terrestrial Implications for Teaching Field Geology Hobbs et al., 1976, p. 347–375) together with work with fi eld computing systems can have a symmetry analyses () direct bearing on how these systems could, and Although we have described some issues and cross-section construction. The fi eld pro- should, be developed for fi eld science on the related to teaching with fi eld computer systems, cedure is time consuming, and in conventional moon, Mars, and beyond. These missions will a number of additional issues are important to paper mapping often produces a map that is present serious challenges that go far beyond emphasize. All students have different ways of completely incomprehensible to anyone but the most hostile environments encountered learning, and in traditional fi eld geology classes the person who made the map. In contrast, by fi eld geologists. Field GIS technology can there is little room for alternative learning tech- fi eld GIS provides the ability to superimpose greatly enable fi eld exploration in hostile plan- niques. That is, there is no other way to collect multiple data layers, turn layers on and off, etary environments where the geologist will be data than to go to the fi eld. During fi eld classes and zoom a fi eld map through a nearly infi - hampered by both terrain and the need to carry there is a widespread “sink or swim” attitude nite scale range, all of which are mapping a life-support system that puts severe constraints that places many students from urban environ- functions that are a remarkable improvement on time and metabolic activity. ments or with physical limitations at a great to anyone who has done this kind of mapping Properly implemented, a fi eld computing disadvantage. Using fi eld computer systems on paper. Furthermore, the recently developed system not unlike the one currently in use by allows a great deal more fl exibility in data col- digital recording compass inclinometer opens us can streamline and simplify the information lection that can benefi t these students in par- options never before considered. These devices resources needed during an extravehicular activ- ticular. Foremost among these capabilities is the can rapidly measure orientations of planes, or ity (EVA), or spacewalk. This can include both potential of 3D visualization systems to permit simultaneously measure the orientation of a the operational checklists required to safely per- virtual access to any place on Earth. Through plane and a line on the plane. We used one of form the EVA as well as the data-gathering tasks, these systems there is tremendous instructional these devices recently and obtained as many as e.g., geologic mapping and sample collection, to potential for both students with no direct access

Geosphere, June 2010 289

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al. near-vertical, oblique view

View down-plunge of folds

Figure 7. Split-screen stereo views of a fi eld area in southern Alaska showing the power of three-dimensional (3D) visualiza- tion with these systems. Yellow lines are bedding traces and red lines are faults. For more details, see the ivs fi les (included herein) of the same data set. Figure was prepared using digital fi les generated during fi eld work in the area, refi ned and edited in ArcGIS using high-resolution imagery and light detection and ranging (LIDAR) topography, draping of the shapefi les onto a LIDAR digital elevation model (DEM) using Fledermaus (see text), and using the Fledermaus viewer for split-screen stereo (e.g., Geowall display). Although this is an elaborate example, similar capabilities can be easily extended to any fi eld project, particularly as 3D display capabilities improve.

290 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

to geologic features near home and for physi- may be a culprit (e.g., see discussion by Clegg a Brunton compass and paper maps as a quaint cally disadvantaged students. Anyone who has et al., 2006). In the future we plan to experiment artifact of the past, reminiscent of the plane taught fi eld geology is familiar with situations with techniques to force continued data synthe- table and alidade for mapping, the blow-pipe for where a student either arrives or becomes physi- sis, but we are uncertain of the ideal approach. mineral analysis, or other archaic technologies. cally incapable of conducting day-to-day fi eld We suggest that fi eld classes should experiment work, and using visualization is one way to deal with different techniques that can aid accelerated APPENDIX 1. NOTES ON with this issue. learning of fi eld techniques, but should recog- HARDWARE SELECTION Another teaching benefi t we have found is nize the variety of learning techniques among the ability to encourage collaborative learning the students. We suspect that as 3D visualization Hardware suitable for fi eld situations depends with the use of the fi eld GIS systems. In a sense, becomes more widely available, the 3D visual- heavily on the use planned for the systems, and for fi eld geology classes have long employed some ization system, together with GPS, will make a the physical environment of the fi eld work. We rec- ognize three key issues for selecting hardware: form of collaborative learning in the use of fi eld powerful accelerated learning tool. Thus, we sug- (1) ruggedization; (2) display; and (3) weight. partners, although working in pairs primarily gest that collectively these are important research serves as a safety factor. Nonetheless, students issues for geoscience educators. Ruggedization working in the fi eld in pairs or in groups have long lead to assessment nightmares because it CONCLUSIONS Although it is tempting to assume that all fi eld work is often unclear whose work ends up in the fi nal requires so-called ruggedized devices, from our expe- result. Field GIS systems do not eliminate all of Our experiences with fi eld GIS mapping rience these devices are generally a waste of money. these problems, but do improve our ability to systems indicate that they are a signifi cant In more than 10 yr of working with these devices we assess individual performance because students improvement over the traditional paper map- have physically broken only two devices and one of can be asked to present their raw data fi les at any ping techniques that have dominated geol- these was a rugged device, which, ironically, cost more to repair than a new nonrugged device. Electron- time during a fi eld project. ogy for more than two centuries. In particular, ics failure independent of accident, however, is more More important, however, using fi eld GIS instrumentation and techniques now exist to common; e.g., in our experience, ~10% of handhelds systems can encourage true cooperative proj- collect large amounts of data rapidly and with undergo a hardware failure within 2 yr. Fortunately, ects and team-building exercises where stu- high precision in the fi eld. More importantly, the consequences of fi eld hardware failures can be minimized by using interchangeable fl ash memory dents can gain familiarity with working in a those data can be merged and viewed with a devices. That is, we keep all data fi les on secure digi- group research environment. For example, variety of other geospatial information, both tal (SD) or compact fl ash cards, which can be removed many fi eld classes have long emphasized a in and out of the fi eld setting. This makes the in the event of a hardware failure. Note, however, that research approach to fi eld problems and used fi eld mapping process much more effi cient and these cards also fail periodically, and thus backup to groups of students to develop regional geologic increases the reliability and repeatability of col- other media is also critical to ensure data integrity. We have found that fl ash card failures can be minimized if maps where student groups were responsible lected data. The ability to combine data sets and fi les are completely rewritten daily to avoid repeated for all steps from data production to data com- create 3D visualizations of mapped structures rewrite cycles on the same sector of the card. There pilation. Although this approach is possible increases the interpretive abilities of the map- are several methods to accomplish this, but the easiest with a traditional paper mapping approach, it is per, both in and out of the fi eld. method is to move each day’s work into new folders, particularly if nightly edits are performed. far easier to implement with a fi eld GIS system. There are some trade-offs that arise when With fi eld GIS, student maps can more easily using computers in mapping. In particular, the Display be merged to produce an aggregated data set user must know something about the software that covers a large area. Students can discuss and hardware as well as implement new proce- and revise the maps as they compile the data, dures to prevent the loss of data in the fi eld. We There remains a complete range of display types available for fi eld devices, but currently the princi- and peer pressure can force improved perfor- have found, however, that we can teach under- pal distinctions in usable outdoor displays is trans- mance when the group is relying on input from graduates, even those with little computer back- missive versus transfl ective displays. Conventional everyone in the group. Thus, use of computer- ground, the necessary skills in a short amount of LCD displays on laptops use a transmissive display, ized systems can encourage a research quality time. We note, though, that in any fi eld effort, and use a backlight to make the screen viewable. In approach in fi eld classes. one or more people must be relatively skilled an outdoor environment, a transmissive display can produce a reasonable display, but usually requires a A second advantage of using these systems at all aspects of the software and hardware to very bright backlight, a situation that produces a trade- for teaching is the accelerated learning of several avoid catastrophic results. Becoming profi cient off between display quality and battery life since the key fi eld skills. One clear example we have seen in using these systems requires a different learn- display produces a large power drain. We have found is, by incorporating GPS positioning for station ing procedure than paper mapping. Everyone is these types of displays acceptable if the main work environment allows access to shade or the work is in descriptions, students more quickly adapt to the familiar with paper and pencil since childhood, cloudy areas. In direct sunlight, none of these types of fi eld techniques of station description and orien- but not everyone is comfortable with using these displays produces a legible display. tation data collection without the classic question high-tech devices. In teaching students the sys- In contrast, transfl ective displays work best in of “where am I?”: unfortunately, it is not clear tems accelerate learning in some areas, but can direct sunlight, and in the shade or on cloudy days, that this step improves performance on other produce obstacles in other areas. For experi- require use of a backlight for clear visibility. As the name implies, these displays have a refl ective LCD mapping skills, because the traditional “where enced fi eld workers the steep learning curve of panel and a subtle issue with these screens is the am I” question also serves to improve map read- adapting to the systems will initially seem an screen surface. Many devices come with a nonrefl ec- ing skills that are essential for things like pro- obstacle, but it is well worth the effort once you tive coating, which improves display dramatically by jecting contacts. Moreover, we have recognized have learned the system. minimizing glare, yet these devices often contain a dilemma; for practical fi eld use, a screen protector is cases where some students experience signifi cant The time has come to abandon paper-based required to avoid destroying surface coatings with the problems with synthesis exercises, and we sus- mapping, and we hope that within a decade pen interface, yet most screen protectors lack nonre- pect the small screen size of handheld computers we will look back on geologic mapping using fl ective coatings. There is currently no solution to this

Geosphere, June 2010 291

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

problem other than to avoid devices that rely heavily APPENDIX 2. GEOLOGIC TUTORIAL because geographic coordinates may be desirable, on screen coatings for glare suppression; although FOR WORKING WITH ARCGIS-ARCPAD- but produce serious map distortions at high latitude. rare, some screen protectors have some glare suppres- ARCPADSTUDIO SYSTEMS FROM ESRI A cautionary note: when reprojecting fi les, be careful sion, although usually at the expense of visibility in with fi le names. ArcGIS defaults to a renaming of the transmitted light (low light conditions). Transfl ective fi les, but this is generally highly undesirable because Although we are advocates of the ArcPad software displays are common on handheld devices, but these of naming issues in the development of the ArcPad system, there are several issues that need to be recog- types of displays are very rare on tablet PCs or lap- layer defi nition fi les (.apl fi les). Thus, it is best to put nized by anyone undertaking using these systems. In tops. Our recommendation, from more than one bad the reprojected fi les into a different folder and either this paper we include two sets of generic shapefi les: experience, is never believe manufacturers advertise- rename back to the defaults during reprojection, or (1) a generic bedrock mapping project that assumes ments of “outdoor display” without seeing the display rename them in ArcCatalog after they have been work predominantly in stratifi ed rocks lacking signifi - in real outdoor conditions. reprojected. If you are not modifying the basic data cant ductile structure; and (2) a generic metamorphic A commonly misunderstood issue with displays that structure or symbology of one of our generic projects structure mapping project where multiple generations is not apparent to most people until they have used a (e.g., you have only added attributes within a given of ductile structure are expected. This pair of projects device is the trade-off between touch-sensitive screens shapefi le, or have added attributes that have no effect versus electromagnetic (EM) digitizers. EM digitizers is not meant to encompass the entire range of projects possible, but serves to represent how projects are orga- on symbology), then you can skip to step 5 (follow- are standard on most tablet PCs and have an advan- ing), otherwise follow steps 2–4. tage in that the screen tracks the pen hovering over the nized and set up in the ESRI system. It also contains script fi les (ArcPad layer defi nition fi les, i.e., fi les with 2. Keep your fi les by themselves in a folder to pre- tablet surface. This feature is extremely useful for cer- vent confusion and errors when converting to ArcPad. tain drawing operations, because linework and mouse “.apl” extension) that show various examples of sym- bol plotting, line style variation based on attribute, and After all of the line and point fi les are in the same pro- emulation operations are relatively intuitive to most jection, start ArcMap and add these fi les to the map. people. Nonetheless, EM digitizers are expensive and labeling. We recommend using these projects initially as a training exercise to get used to using the software, Once the fi les are in ArcMap, you should begin by set- are virtually unknown on handheld devices. Instead, ting up the linework for the map, and the line styles these devices typically use a pressure-sensitive touch but after using the systems it is virtually certain that users will need to customize the projects to their needs. you build in ArcMap will be exported to ArcPad. Typi- screen system. In touch screens, screen calibration is cally the linework will be drawn either generically by critical, as well as sensitivity. For example, there are It is possible to develop a project entirely in ArcPad, but the function of the system will be limited. To fully layer (e.g., all faults are red, heavy lines) or based on few things more annoying than an excessively sen- some attribute (e.g., solid line for exposed contacts, sitive touch screen because something as simple as customize the system, you will need two additional pieces of software: (1) ArcGIS (an Arcview license dashed line for fl oat contact). Note that if you are laying the palm of your hand on the screen can make building the project as a blank project, and plan to the pen jump to your palm. Fortunately, most devices is suffi cient) and (2) ArcPad Studio. We have wasted days of effort learning the range of software bugs and draw linework based on an attribute, you must begin can be adjusted for these issues, and modern devices by generating one line for each attribute you plan to rarely have this problem. However, readers should quirks in the software that are required to build a rea- sonable project. From that experience, we emphasize use for different line styles and fi lling the attribute use care with some touch screens now appearing on fi eld with each attribute. In our examples, the fi eld many smartphones, that are heat activated and will not several key concepts here for people fi rst attempting to build their own project. It is particularly important “contactquality” is fi lled by the attributes exposed, work with a stylus; these devices are useless for fi eld exposed_projected, approx_fl oat, and so forth, and mapping applications where high precision is needed. to follow a specifi c workfl ow in setting up or modify- ing an ArcPad project. In particular, the sequence of different symbology is attached to each line. This step Touch screens are also used on some windows PCs needs to be repeated for every line object in the map; and these systems also have a hidden pitfall related starting the project in ArcGIS and completing it in Arc- Pad Studio is critical; the reverse procedure (starting e.g., in our generic mapping shapefi les, this includes to mouse clicks. With a pen interface, a single tap is depositional contacts, faults, unconformities, fold typically used to simulate a mouse click and double in ArcPad and ArcPad Studio) produces unpredictable outcomes. This workfl ow recommendation is not in axial traces, and intrusive contacts. Regardless, the tap a double click, but there is no equivalent to a right- key is to develop the linework that best fi ts individual click. EM digitizers typically contain a special button any ESRI documentation that we have seen, and the ArcPad to ArcGIS sequence was not even anticipated need. A cautionary note: linetypes can be changed in to accommodate this issue, but for touch screens this the fi eld in ArcPad, but results can be unpredictable; is usually dealt with by a special button on the device, by ESRI based on correspondences we had with the company during some of our work. Thus, we advise thus, we recommend avoiding those fi eld changes a situation that can lead to awkward hand positions for unless absolutely necessary. some operations. following this sequence. 1. Unless you are building your project entirely 3. After developing a set of line defi nitions in Arc- in ArcPad, begin building your project in ArcGIS. If Map, work on symbology of point fi les. For simple Weight and Size you are building the project from scratch, you should point-based fi les like stations, this operation is sim- fi rst create a geodatabase and develop all of your line ply choosing a symbol, but for orientation symbols We indicated briefl y that weight is a large factor and point objects within a single feature class, and this procedure is complex. Indeed, it is suffi ciently for anyone carrying these devices while doing fi eld export the project from the geodatabase after the steps complex that we highly recommend starting with our work on foot. What constitutes an acceptable weight described below. Geotadabases and shapefi les are cre- “orientation” shapefi le before attempting to build your depends on the person, but we have generally found ated in the ArcCatalog program. Usually, a personal own variant. If you do choose to build a variant from that any device >2 kg will generally get left behind by geodatabase is the best option. New feature classes scratch, the following are the key steps. fi eld workers. Similarly, form factor of the device is are created in the geodatabase, and given common A. Choose an appropriate symbol based on attri- an issue, but depends greatly on the trade-off between projections. When creating a new feature class, be bute (chosen in the symbology tab of the layer proper- screen size and portability. That is, it is doubtful many sure to check the Z-values box to record elevations. ties) by right-clicking the selected layer, followed by people would carry a 30 in (76 cm) tablet in the fi eld, This will be critical if you plan to do any 3D work. choosing “unique values” in the categories fi eld. but a notebook-size tablet would be acceptable to Then the desired fi elds can be added. For an easier B. Use the pull-down menu under the “value” many; others would prefer a device that could be car- time, import the fi elds from the shapefi les provided fi eld in the categories window, then click the “add ried in a vest or hip pouch. with this report and modify or add to them. all values” button. As with linework, if the project Shapefi les are generally easier to use, however, is blank, you will need to have one point with each attribute you intend to sort on as “initial” data to Summary because they are the native format for ArcPad. If you plan to use the shapefi les provided here as templates, undertake this step. take those fi les and begin by reprojecting them to C. At this stage, change the symbology to your In any case, the key issue for anyone is that it is the coordinate system you plan to use. ArcPad can- choices (e.g., a standard bedding strike and dip sym- very important to test any device planned for fi eld not pro ject fi les on the fl y, like ArcGIS, and so all bol versus foliation), and when all symbology has operations. All devices ultimately require a compro- fi les must be in the same coordinates. Thus, some been chosen, complete the operation by going to mise, and the ideal device for a given user will depend initial planning can save time. Project builders must the “advanced” button in the lower-right corner of strongly on the applications and environments where decide if it is easier to reproject all fi les into a uni- the symbology tab window. In the pull-down menu, the devices will be used. We generally prefer nonrug- versal reference frame like WGS84 (i.e., World Geo- choose “rotation” and a pop-up dialog box appears. In gedized, smaller form-factor devices (UMPCs and detic System), or is it easier to reproject later and the dialog box, choose the “geographic” button, and handhelds) with transfl ective displays for fi eld work, work in an odd reference frame like NAD27 (i.e., in the pull-down menu, choose the attribute fi eld that but others clearly prefer alternatives. North American Datum)? Map projection is an issue corresponds to the azimuth fi eld for rotation of the

292 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Field computer systems

symbol: typically strike for planar features recorded for building a custom form and is well explained in Insulated wire, ~30 m, ~#18 usually suffi cient, 2 with a right-hand rule. the documentation. Cautionary note 1: Although Arc- color-coded single-strand wire is best D. Go to the labels tab and select the fi eld for dip Pad Studio is relatively intuitive for building custom Cigarette lighter accessories (highly desirable, both measurements as a label fi eld. forms, the program does not always generate a reli- for car charging systems and solar systems for quick E. A cautionary note: for those accustomed to dip able script. The program frequently generates ghost connections; remember, however, that the center ter- direction and dip measurements or strike and dip mea- boxes that do not show up in the “test form” fi eld of minal is always + on these systems) surement without the right-hand rule convention, there the program, but do appear when the script is run in Electrical tape is as yet no workaround for recording data this way. ArcPad. The best way to fi nd these ghost fi elds is to Screw drivers (+ and -) Similarly, some default symbols will not work with look at the raw script fi les, searching for items like Wire nuts this method; e.g., fold axis symbols that are preloaded a combo box or edit box with no attributes named in Miscellaneous electrical connectors and/or termi- in ArcGIS are 90° off. Note that if you are planning to the script. This is most easily done by comparing an nal blocks(optional) export the project to a handheld device, the handheld anomalous script to a working script. Cautionary note device must have the custom ESRI fonts installed for 2: ArcPad Studio has a bug that has yet to be fi xed as Check List for Field Projects the symbology to plot properly. The easiest method of version 7.1. In combo boxes or list boxes, the pro- is to use a PC, navigate to the font folder in the Win- gram offers two fi elds for fi lling attributes in the pick For each device being used (repeat checklist for dows directory, and fi nd all of the ESRI fonts. Copy list. This implies that one of the fi elds is a label and each device): these fonts to the clipboard, and paste them into the the other is the actual text that will fi ll the fi eld in the Device 1 font folder in the Windows directory of the Windows attribute table. In fact, only one of these fi elds is actu- Software loaded and operational mobile device (typically this will be done through the ally written to the data table. Thus, it is important to Project and data loaded Active-Sync application from Microsoft). make both fi elds the same, even though it is tempting GPS operation tested (this may require a simple F. When all symbology is set, the data need to be to make more easily understood labels in the forms. test, or a more elaborate test with wireless devices like exported to ArcPad. This process seems straightfor- In addition, combo boxes or list boxes can also call a Bluetooth GPS units) ward, but there are bugs in the export software with .dbf table containing the attributes for that fi eld. This Memory card installed and data paths set to write versions as recent as ArcPad 7.1 that produce unpre- feature is highly desirable for fi elds that are likely to to memory card dictable results. A known procedure that works is fi rst, change routinely; e.g., we use a fi eld for “geologist” in from the ArcMap window, right click on each shape- our station table to identify the person who collected Other Equipment Needs fi le for which you have developed symbology, and the data, and it is easier to make this change in a .dbf scroll down to “save as layer fi le”; this will create a table than constantly revising the .apl fi le. Unfortu- dialog with a fi le with the same name and a .lyr exten- nately, this requires some care. Specifi cally, it is very 110 V power available sion, or a layer defi nition fi le. Save each layer defi ni- important to remember to always copy this .dbf table Chargers for all computers packed tion in the same fi les as the shapefi les. Second, open a when generating a new project from an old one. 12 V inverter needed for any equipment (if this new ArcMap window and add all of the layer defi ni- Moreover, we have had unpredictable problems equipment is needed, you may need a separate check tion fi les you just created. The result should look iden- with .dbf table links being broken or corrupted, pro- list for cigarette lighter accessories and other items) tical to your other ArcMap window: if it doesn’t, fi x ducing variable results. Thus, until some of these Power strips suffi cient for charging all devices the parts that did not export properly, or fi x the links, problems are fi xed, we tend to avoid using .dbf tables 12 V system with stand-alone battery and solar and repeat. When this fi rst export is complete, mini- for attributes, except in cases where the process is charging: mize or close the original ArcMap fi le. Third, add all highly repetitive; e.g., several shapefi les using the All wiring and connection system in place and of the remaining data you wish to use in your ArcPad same set of attributes. tested? project to the newly opened map. This typically will 110 V inverter(s) tested and operational? include mostly rasters such as aerial photographs, dig- Spare fuses? ital raster graphics (DRGs), and shaded reliefs from Power needs consistent with equipment on hand APPENDIX 3. FIELD EQUIPMENT CHECK (e.g., inverters, battery capacity) DEMs, but may also include other vector fi les. Make LIST AND SPECIALIZED EQUIPMENT sure all of these fi les are in the same projection before Tool kit packed proceeding. If ArcMap is projecting them on the fl y, Battery tested for voltage and power (this can be done at most auto supply stores, or you can buy your reproject them to the map reference frame, add the In a fi eld project involving computers we suggest own tester. Necessary if you are using an old battery) reprojected map, and remove the fi le that is not in the use of a checklists and carrying specialized equip- Solar panel tested same projection. A cautionary note: ArcMap does not ment for troubleshooting. This type of procedure is 12 V system using auto charging system tell you what projection it is using when using on the probably familiar to most geophysicists, but is alien fl y projection, and you can be easily fooled if you do to most geologists. The lists here are only an exam- not check. The program defaults to the projection of ple, and would generally need specialization to spe- ACKNOWLEDGMENTS the fi rst projection it recognizes when fi les are added, cifi c applications. even if the fi rst fi le is later removed! Thus, to avoid This paper would not have been possible with- problems, if any fi les are in a different projection, it is Specialized Equipment for Remote Sites Using out the generous support of a series of research and best to exit ArcMap, restart the program, and remake Solar Charging Systems or Car Chargers equipment grants that allowed us to test various a base map with a layer defi nition known to contain types of equipment. These include our fi rst tests of the correct projection. Fourth, use the ArcPad Data equipment supported by National Science Founda- Manager to export the map features to make an Arc- 12 V battery (size dependent on project needs; tion (NSF) grant EAR-9706233 to Pavlis and Serpa Pad project. This tool does not automatically appear in minimum of 2–3 Ah/user) and the Louisiana Board of Regents (BOR) Support ArcGIS and is added from the Tools menu of ArcMap Electrical multimeter (DC volt, ohm, continuity fund (1998); a Keck Foundation grant to the Univer- under “customize.” If you have ArcPad 7.1, do not use minimum; DC amps useful) sity of New Orleans, which supported our fi rst major the generic ArcPad tools under the customize tab in Solar panel (20–40 W typical for small group) equipment purchase; a second Louisiana BOR Sup- ArcPad 7.1 as this tool is a legacy of earlier ArcPad Note: to calculate power requirements, recall port grant (2003) supporting equipment update; and a versions and is unreliable in ArcPad 7.1, although it is Joule’s law: P = IV. A solar panel normally outputs University of Texas (UT) STARS grant to UT El Paso the only tool in earlier versions. 18 V, but must be regulated to 12 V for battery charg- that supported purchase of our present generation of 4. If you are using the fi les provided here as tem- ing. Thus, for a 40 W panel its full-sun output will be equipment. Research grants that used this equipment plates, they should work in ArcPad 7 or 8. How- at 2–3 A. With 4 users, using 2-3 Ah/day, this illus- and helped support the effort include NSF grants ever, you may want to change the look of the forms trates that ~4 h of full sun will be needed to main- EAR-9706233, EAR-9725035, EAR-9910899, EAR- we have built. If so, or if you are building your own tain a fully charged battery. This arithmetic can be 229939, EAR-0409009, EAR-0735402, and EAR- project, you will now need to open the .apl fi le that adjusted for any fi eld area, but it is critical to build in 0711105. We thank Midland Valley Ltd. for access to was exported from ArcMap in ArcPad Studio, and at least 2–3 days of spare power for bad weather, or Move software through academic discounts and for build custom forms. This is accomplished by pulling other eventualities. assistance in using the software. Development of our down the forms menu, then click on “edit form” in Wire cutters present system would not have been possible with- the popup menu. The interface is a graphical interface Needle-nose pliers out input from Boyan Brodaric, Evan Thoms, Doug

Geosphere, June 2010 293

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021 Pavlis et al.

Walker, and George Brimhall, as well as collaborative mapping: Computers & Geosciences, v. 30, p. 5–20, Pavlis, T.L., 1999, Ramblings of a dusty fi eld geologist: The researchers in the St. Elias Erosion and tectonic Proj- doi: 10.1016/j.cageo.2003.08.009. pros and cons of high-tech fi eld tools: Geological Soci- ect (STEEP) and students working on various research Brodaric, B., Gahegan, M., and Harrap, R., 2004, The art ety of America Abstracts with Programs, v. 31, no. 7, projects listed here. We thank David Mogk for helpful and science of mapping: Computing geological catego- p. A192. ries from fi eld data: Computers & Geosciences, v. 30, Pavlis, T.L., and Little, J., 2001, Using handheld personal comments on improving the manuscript. p. 719–740. computers as fi eld data collection tools: Some lessons Bruhn, R.L., Pavlis, T.L., Plafker, G., and Serpa, L., 2004, learned in the school of hard knocks in the Wingate REFERENCES CITED Deformation during accretion in the Saint Elias Wash project and related projects using Fieldlog/Field- orogen, Alaska: Geological Society of America Bulle- worker software exported to Arcinfo, in Soller, D.R., tin, v. 116, p. 771–787, doi: 10.1130/B25182.1. ed., Digital mapping techniques ’01—Workshop pro- Asch, K., 2003, Digital map : No more hiding Campbell, E., Duncan, I., and Hibbits, H., 2005, Analysis of ceedings: U.S. Geological Survey Open-File Report places for inconsistent geologists: Geologija, v. 46, errors occurring in the transfer of geologic point data 01-223, p. 115–122. p. 329–323. from fi eld maps to digital data sets: U.S. Geological Pavlis, T.L., and Sisson, V.B., 2003, Development of a sub- Asch, K., 2005, The new digital geological map of Europe Survey Open-File Report 2005-1428, http://pubs.usgs horizontal decoupling horizon in a transpressional and standarisation: Consistency as the last refuge of the .gov/of/2005/1428/campbell/index.html. system, Chugach Metamorphic Complex, Alaska: unimaginative?!, in Ostafi czuk, S.R., ed., The current Clegg, P., Bruciatelli, L., Domingos, F., Jones, R.R., DeDo- Evidence for rheological stratifi cation of the crust, in role of geological mapping in geosciences: Proceedings natis, M., and Wilson, R.W., 2006, Digital geological Sisson, V.P., et al., eds., Geology of a transpressional of the NATO Advanced Research Workshop on Inno- mapping with tablet PC and PDQA: A comparison: orogen developed during ridge-trench interaction along vative Applications of GIS in Geological Cartography: Computers & Geosciences, v. 32, p. 1682–1698, doi: the North Pacifi c margin: Geological Society of Amer- NATO Science Series IV: Earth and Environmental Sci- 10.1016/j.cageo.2006.03.007. ica Special Paper 371, p. 191–216. ences Volume 56: Dordrecht, Springer, p. 1–9. Ernst, G., 2006, Geologic mapping—Where the rubber Pavlis, T.L., Picornell, C., Serpa, L., Bruhn, R.L., and Black, R., and Walker, J.D., 2001, Development and use of a meets the road, in Manduca, C.A., and Mogk, D.W., Plafker, G., 2004, Tectonic processes during oblique- laptop-based geological mapping system: Experiences eds., Earth and mind: How geologists think and learn collision: Insights from the St. Elias Orogen, northern at the University of Kansas, in Soller, D.R., ed., Digi- about the Earth: Geological Society of America Special North American Cordillera: , v. 23, TC3001, tal mapping techniques ’01—Workshop proceedings: Paper 413, p. 13–28, doi: 10.1130/2006.2413(02). doi: 10.1029/2003TC001557. U.S. Geological Survey Open-File Report 01-223, Golding-Luckow, H., Pavlis, T.L., Serpa, L., Guest, B., Wag- Phelps, G.A., Boucher, A., Jachens, R.C., and Simpson, p. 127–131. ner, D., Snee, L., Hensley, T., and Korjenkov, A., 2005, R.W., 2009, Constructing a 3D geologic map: Geologi- Brimhall, G.H., and Vanegas, A., 2001, Removing science Late Cenozoic sedimentation and volcanism during cal Society of America Abstracts with Programs, v. 41, workfl ow barriers to adoption of digital geological transtensional deformation in Wingate Wash and the no. 7, p. 37. mapping by using the GeoMapper Universal Program Owlshead Mountains, Death Valley: Earth-Science Riggs, E.M., Leider, C.C., and Bailliet, R., 2009, Geologic and visual user interface, in Soller, D.R., ed., Digital Reviews, v. 73, p. 177–220, doi: 10.1016/j.earscirev problem solving in the fi eld: Analysis of fi eld naviga- mapping techniques ’01—Workshop proceedings: U.S. .2005.07.013. tion and mapping by advanced undergraduates: Journal Geological Survey Open-File Report 01-223, p. 103– Guest, B.G., Pavlis, T.L., Serpa, L., and Golding, H., 2003, of Geoscience Education, v. 57, 15 p. 114, http://pubs.usgs.gov/of/2001/of01. Chasing the Garlock: A study of tectonic response Thoms, E.E., and Haugerud, R.A., 2006, GDA (Geologic Brimhall, G.H., Vanegas, A., and Derek Lerch, A., 2002, to vertical axis rotation: Geology, v. 31, p. 553–556, Data Assistant), an ArcPad extension for geologic GeoMapper Program for paperless fi eld mapping doi: 10.1130/0091-7613(2003)031<0553:CTGASO> mapping; code, prerequisites, and instructions: U.S. with seamless map production in ESRI ArcMap and 2.0.CO;2. Geological Survey Open-File Report 2006-1097, GeoLogger for drill-hole data capture: Applications Haugerud, R.A., Richard, S.M., Soller, D.R., and Thoms, http://pubs.usgs.gov/of/2006/1097/. in geology, astronomy, environmental remediation E.E., 2009, NCGMP09—A database schema for digi- Thoms, E.E., Soller, D.R., Haegerud, R.A., and Richard, S., and raised relief models, in Soller, D.R., ed., Digital tal publication of geologic maps: Geological Society of 2009, Database schema for NCGMP09—A proposed mapping techniques ’01—Workshop proceedings: America Abstracts with Programs, v. 41, no. 1, p. 38. standard format for digital publication of geologic U.S. Geological Survey Open-File Report 01-223, Hobbs, B.E., Means, W.D., and Williams, P.F., 1976, An maps: Geological Society of America Abstracts with p. 141–151, http://pubs.usgs.gov/of/2002/of02-370/ outline of : New York, John Wiley & Programs, v. 41, no. 7, p. 279. brimhall.html. Sons, 571 p. Walker, J.D., and Black, R.A., 2000, Mapping the outcrop: Brimhall, G.H., Dilles, J., and Proffett, J., 2006, The role of Lipovsky, P.S., Colpron, M., Stronghill, G., and Pigage, L., Geotimes, v. 45, no. 11, p. 28–31, http://www.geotimes geological mapping in mineral exploration, in Doggett, 2003, Geofi eld v.2.2—Data management and map pro- .org/nov00/mapping.html. M.D., and Parry, J.R., eds., Wealth creation in the min- duction for the fi eld geologist: Yukon Geological Sur- erals industry: Integrating science, business and educa- vey Open-File 2003–8(D). tion: Society of Economic Geologists Special Publica- Mathers, S., Kessler, H., Wood, B., and Sobisch, H., 2009, tion 12, p. 221–241. The deployment of 3D geological modeling at the Brit- MANUSCRIPT RECEIVED 10 FEBRUARY 2009 Brodaric, B., 2004, The design of GSC Fieldlog: Ontology- ish Geological Survey: Geological Society of America REVISED MANUSCRIPT RECEIVED 03 NOVEMBER 2009 based software for computer aided geological fi eld Abstracts with Programs, v. 41, no. 7, p. 38. MANUSCRIPT ACCEPTED 28 DECEMBER 2009

294 Geosphere, June 2010

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/6/3/275/3339183/275.pdf by guest on 02 October 2021