On field survey methods using handheld GPS

The previous post dealt with the various software packages we use for desktop mapping, ArcView, ArcCatalog etc. In this post we will look at the software we use to complete mapping tasks in the field.

ArcPad is a mobile version of ArcView and can perform many of the same functions as its desktop counterpart, but is designed for handheld devices used in the field. The main benefit in using ArcPad is the ability to combine map data with a GPS receiver on the mobile device. The GPS receiver, through various conversions, communicates with ArcPad to plot the location of the Trimble (and thus the person holding it) on the screen, relating it to the relevant map data.

After a few minutes to allow the GPS receiver to pick up a sufficient number of satellites, a crosshair appears on the screen indicating our location. As we work, the receiver will track our movements in real time and ArcPad will update our geographic position on the screen map, including point locations and shape files we have added. Thus, we can use the combination of the GPS position and the locations of the historic lime mines (stored in the handheld’s memory in Vector format) to guide us to the location of a specific mine, within an error range of a kilometre or so.

Image of ArcPad using GPS to plot a current position (crosshairs) on the Trimble Handheld. The points represent either historic mines or sites added to our database, depending on what colour they are.

Not only can we use existing files in the field, ArcPad also allows us to edit blank shapefiles and add data for new locations in the field, along with text-based attributes to describe relevant features of each location. This is another great advantage to mobile GIS. As explained in the previous post we use ArcCatalog to create blank shapefiles. Once these are loaded onto the handheld and opened in ArcPad we can use the GPS receiver to record our location and store it within the shapefile. This is how we add new locations to our database and record a more accurate location for any historic lime mines we find.

A major downside to using handheld platforms is that they do not have particularly quick processors. As such, using large raster data sets, like topographical maps, is tricky. It is possible to convert raster data to a size that ArcPad can handle, but this has limitations on the amount of data it can convert. Consequently only a very small section of a topographical map can be converted for use on the handheld. For survey work over large regions, this is not really useful.

Given the large scale of the project it has not been feasible, or deemed necessary, to perform this conversion over and over again for each of the regions we have visited. Rather we have mainly focused on using the digitised vector data to represent point locations and geological units extracted from the mining and geological maps for use on the handheld. This takes quite a bit of time to prepare but has provided invaluable information to us in the field. Once we have finished surveying for the day we download the shapefiles with the new location data on them onto our tablet PC and add them as new layers to ArcView. We can now view the points along with the topographical map, enabling us to view the new information in a wider context.

To summarise, ArcPad allows us to:

  • Know our exact location through the GPS receiver
  • View pre-existing Vector shapefiles like our historic mine location database and geological formations
  • Add new data blank Vector shapefiles with points or polygons captured using the GPS receiver
  • Upload the new information to our tablet PC viewing it in a wider context and with other relevant data like topographic maps

Without ArcPad and the GPS function of our handhelds relocating historic lime mines and plotting the accurate location of new sites for our database would be incredibly difficult. Using ArcPad and ArcGIS for mapping also has the added advantage of being able to connect directly to our project database in Microsoft Access and update our field data on daily basis. The next post on our survey procedures will deal with the database software and how we use both desktop and handheld-based software to record the sites we visit, and how all the components work together.

– A Reid

We think we found a fossil site – now what?

When we locate a potential site locality, we begin a process of exploration and evaluation to determine whether there is potential for further research at that locality. Our basic tools are a handheld GPS, a rock hammer, and a panga (or machete).

Field crew, 2nd trip: Andy with the GPS, Morris with the rock hammer, Kevin with the machete: Did the timer go off? is the camera working? did anyone hear a ‘click’?

The main objective of our project is to locate historic mining sites, and to evaluate whether they contain deposits and other features that are useful to us as palaeoanthropologists. We are looking for a variety of in situ (i.e., undisturbed) deposits including:

  • speleothem or cave formations, which can be sampled to obtain geological dates, and for stable isotopes which tell us about past climatic and environmental conditions;
  • cave infill sediments (derived from soil and other debris from both outside and inside the cave), which can tell us how the cave filled in, and also may contain evidence of the outside environment such as pollen;
  • fossil bone, including those of reptiles, birds, and rodents, as well as of larger animals including bovids, carnivores, primates, and of course, our own ancestors (hominids).

For some sites, the presence of such evidence can be quite obvious, especially if there are dense fossil deposits just waiting to be excavated. However, at other sites, it can take considerable effort to determine whether this evidence is present. Some sites are small and the deposits easily located, and others – especially when the mining activity was extensive – can be complex and more difficult to define.

One of the underlying objectives of this project is to consider what evidence at such sites can be used to answer questions of interest to palaeoanthropologists even in the absence of fossils. After all, when we find fossils (and especially those of hominids), we then use a battery of sampling techniques such as those above to obtain information about the temporal, environmental, and depositional context of that fossil. But since fossils are relatively rare in the geological record, we may be missing out on valuable information about past environmental conditions by passing over deposits in which hominid or other fossils are rare or even absent. This information could help us to refine our models about environmental and adaptive change during hominid evolution.

So, after that extensive preamble, what do we do when we think we have located a site? There are three basic tasks to complete: defining the limits and features of the site; mapping the locality; and evaluating the site’s potential for future research.

Defining the limits of the site involves exploration of the site’s geographical and geological features. We walk up and down hills, climb rock outcrops, and (carefully!) climb into mine shafts and trenches, looking for speleothem deposits, breccia and cave infill sediments, and fossils.

A miner’s trench exposing a lot of geological formation: Can anyone see any breccia up there?

One of the challenges in this work is being able to see the ground clearly in order to find features, breccias blocks, and of course, fossils. This is where the panga comes in handy!

Like most arid regions, many of the plants here have thorns. This site is overgrown with acacia trees and brambles. Keep repeating to self: ‘This is better than being stuck behind my desk at the universiity!’ (There is actually a nice cave infill deposit in this view… just to the right of the machete!)

The miner’s dumps are a great source of information about the deposits on the site. We comb the hills of rock, searching for fossils and other evidence of the nature of the underground deposits. Depending on what we find, we then have to search the mine trenches and chambers to determine where the blocks in the dump came from.

Fossils can be difficult to ID in the field. Observe the fossil braincase (below the knife to the right), and a vertebra (to the right on the edge of the block) of un-named mammals in this breccia block.

Finally, all of the features of the site are mapped using the handheld GPS, which locates each data point taken to within approximately 5 meters. For our initial fieldwork survey, this is good enough, but if we were to return for sampling or excavation we would use total station survey equipment and map the site with more accuracy.

Andy enters data points on the Trimble GPS at a small locality near Gondolin, North West Province.

The initial survey and exploration of a new locality can take several days, and also includes extensive photographic and written documentation; we also record information about the site on databases on the handheld GPS and a table PC. The final task is to sit in the shade and discuss what we have observed on site, and evaluate the potential for further work in the form of sampling or excavation.

Once we have completed this process of site survey and evaluation, we go back to the maps, head out to a new area, and begin the process again. All in a day’s work!

– KL Kuykendall

Geology maps: where is that dolomite?

Apart from the two historic mining maps we have acquired, a substantial amount of our information has come from 1:250,000 maps of the bedrock geology. These provide the geological characteristics, and the location, of the underlying rocks in any area of South Africa. Because the types of caves we are looking for form in dolomite, locating areas where this rock is present is one of the main factors in determining our survey strategy.

The geology maps have been treated in the same way as the historic mining maps, utilising only the information relevant to the project and converting that information into a usable digital format. However, unlike the mining maps, on which the digital information is characterised by a point, the information we require from the geology maps relates to the region (including the size and shape) occupied by a lithological unit of bedrock.

Geology base map. An example section of a 1:250000 geological map showing areas of dolomite in different shades of blue (map published byDepartment of Mineral And Energy Affairs, Republic of South Africa 1981).


In order to convert this map information from Raster to Vector formats, we trace around the edges of the bedrock formations we are interested in, save them as a polygon (instead of a point), and then add the lithological information as an attribute of the feature. Not only does this kind of digitisation help in establishing areas to investigate but it also has a number of other advantages.

Geology map with polygons.  In this figure, the blue areas denoting a dolomitic lithological unit have been overlain by a traced polygon using ArcGIS and are now represented by different colours to indicate different attributes the rock unit might have, such as including interbedded chert, quartzite, or shale (map published byDepartment of Mineral And Energy Affairs, Republic of South Africa 1981).

The first advantage is the reduction in size of the file from a large scanned map which may be as much as 390 Megabytes per file to a few Kilobytes of data. This is helpful not only when trying to use the data on a handheld computer but also for storage on the device, since we can fit many more polygons than scanned maps on a 4 Gigabyte memory card. The second advantage of digitising in this way is that we can select the only data that are  relevant to the project and its aims. We can ignore all the other geological formations which do not hold the appropriate characteristics and only digitise those that we need. This again saves space on the memory cards but also saves time, both in the tracing of the geology and in locating regions appropriate to the project in the field.

Geology polygons. This figure shows how the digitisation of the geological formation looks when the area is loaded onto the handheld computers. The pink and orange colours represent the (dolomite) geology and the green background represents the Vector outline of the main provinces of South Africa.

The “Identify” function on the GIS software (ArcPad) we use on the hand-held computers can be used to call up the attributes of any of the Vector layers  once defined. The GPS function on the handheld also enables us to determine our current field position relative to the geological formations (polygons) and to the mines (points). This helps us to avoid spending a lot of time driving around searching in the field!

This is the final step in the digitisation process to prepare the maps for the field. The next blog in this series will deal with some of the hardware and software we are using in more detail.

– A Reid

Map accuracy: are we there yet?

The project is focused on the relocation of historic lime mines sites across South Africa. Information we have gathered includes high resolution scans of two maps showing the location of every mine in South Africa open at the time of printing. It is from these maps, along with other documentary sources, that we have selected the areas to visit. As you can see below, the location of the mines is clearly marked with a point and a two letter code relating to the mineral being mined with Ls indicating lime (on a side note I have been keeping a beady eye out for those marked as Diamond and Gold, you never know what they might have left behind!).

© Department of Mines, Republic of South Africa 1959

Given that the maps are of a relatively large size it was not feasible to bring them into the field, which even if we did would prove difficult to glean any particularly useful location data from anyway. The solution to this is to make a digital version of the map from the high-resolution scans. Unfortunately the handheld computers we are using do not have a particularly large processing capacity so loading an A0 map scanned at 300dpi and then trying to use it, isn’t efficient or effective. So, what we have to do is to store the data as a much smaller file so it can be used on the handheld computers. This transformation is quite common when using maps like these in an archaeological context, digitising vector data (lines and dots) from raster data (pictures and photos etc.). Vector data files are much smaller than raster data files and they can hold text information about them which is very helpful in the field and saves a lot of time flicking through notebooks.

By using a GIS (Geographical Information System) package it is possible to create a file displaying the locations of the mines as dots, which we can then load onto our handheld computers and use its GPS function to guide us to them. Oh so simple I hear you say, alas no such luck I’m afraid.

A selection of digitised mine locations as displayed in vector format overlaid on a 1:50000 topographical map in raster format.

The main issue with the maps is their accuracy, and furthermore the accuracy of the points we take from the maps. Firstly it is difficult to say how accurately the cartographer plotted the locations. Secondly the scale of the map is so large that the effective accuracy of any point will only be within 1.5 kilometres, in any direction, of the actual location of the mine.  So, even when we are standing right on top of the point stored on the handheld, the mine could be anywhere in a 9 square kilometre area. That’s where we depart from our GPS system to locate the mines and use our eyes, and feet, to find them (always being careful to avoid falling into sinkholes and the occasional thorny bush).

This was only the first step in our preparation for fieldwork, the next will be covered in a subsequent blog post and will deal with how we use and prepare geological maps for the project.

-AR Reid