Banner
Laser technology reveals more of the Maya world in Belize Print E-mail
( 0 Votes )
Written by Administrator   
Thursday, 01 March 2018 00:00

A new laser technology is revolutionizing archaeology in Belize and the Americas. Now four experts in the field, Dr. Anabel Ford, Thomas Crimmel, Claudia Knudson, Sherman Horn and Keith Clarke have sent a joint article to the Guardian Newspaper entitled, ‘LiDAR is not a Magic Wand’, which reveals some of the breakthroughs in Belize, notably at El Pilar, an archaeological reserve that straddles the Belize-Guatemala border.

As their joint article notes, “...it has been called a revolution for archaeology.” Sensationalized for seeing through the forest canopy, and touted for its remarkable imagery that maps the forest floor, LiDAR has captured the attention of the public even as the Maya forest still keeps some of its secrets.  The news of the extensive LiDAR coverage in the “Reserva de la Biosfera Maya” is indeed an important step towards understanding the nature of the ancient Maya landscape.  This extension in Guatemala is matched in western Belize from Caracol north to city El Pilar, located on the border between Belize and Guatemala, making for an amazing source for new research on Maya settlement. The field work to validate these data will undoubtedly take decades to unpack. Combining the Belize and Guatemala LiDAR coverages, we have a total area of some 3000 km2, representing the greater Petén that embraces the northern Guatemala and western Belize, the core area of Classic Maya civilization (Figure 1 ).

This new LiDAR coverage grants the ability to identify every major Maya monument, the “big stuff,” and the claims of 60,000 structures, roughly 30-50 structures/km2, forms a reasonable base to initiate field validation. At the El Pilar Archaeological Reserve for Maya Flora and Fauna, where experts have been working with LiDAR coverage for four years, and find this technology is a magnificent tool (Figure 2).  The data on topography alone is awesome, and can map the variations in elevations to 1m with phenomenal details of the landscape. BUT: LiDAR is not a magic wand.  There is still a vital need to use ground-truthing procedures to validate findings by archaeologists in the field.

Working with our El Pilar LiDAR coverage, a 20 km2 area that straddles the international boundary of Belize and Guatemala (Figure 3),  Dr. Anabel Ford and her  team started by looking at the major monuments that were indeed startlingly clear! But the small structures and other cultural features, such as depressions, terraces, and quarries, were far subtler. From LiDAR imagery, they identified all the elements that were potentially cultural remains—a total of 1,621 “GoTo” points . Based on these identifications, they have field validated everything within a 10 km2 area, including the immediate surroundings of the Late Classic core of El Pilar: 1,214 points in total.  They have found an average of 83% of our points of interest result in cultural features on the ground, but only 611, or 50% of the “GoTo” points, are attributed to structures and mapped as residential units of the ancient Maya people.

Variability is evident across space. The NE area of the El Pilar Archaeological Reserve is a transitional wetland where they have tabulated few elements: 133 “GoTo” points, or 53/km2.  Uncertain of the accuracy of the LiDAR with so few targets for our “GoTo” points,   determined to conduct a full coverage survey in 2016, as  they would have done along survey transects prior to LiDAR-based archaeology.  The group of experts were able to verify the accuracy of the LiDAR imagery in the field by validating the few cultural remains in the area (127 features, including 80 residential units). The majority of the sites, however, were located in the slightly higher southern portion of the area, which rises out of the wetlands and is near a minor center, Chorro, approximately 2 km from the El Pilar core next to a spring of the same name. It is noteworthy that 5% of the “GoTo” points in this area were rejected, and these points were attributed to debris of the corozo or cohune palm, Atelea cohune (nuts and dead palm fronds) that accumulate beneath these majestic palms. Curiously, there were unusual linear features recorded on the ground in the wetland area that were not visible in LiDAR images These alignments are at contours, suggesting water control features, but then again, they could be something else.

In the NW, where large rectangular features were identified in LiDAR images, 317 “GoTo” points, or 120/km2, were identified.  Of the 317 “GoTo” points, 299 (95%) revealed cultural features, including quarries, depressions, terraces, berms, and chultunes, as well as 150 residential units, all recorded in the 2017 survey.  Only 5% of our points were rejected as other types of features, which were most commonly large tree buttresses, such as found on amate trees.  Large plazas, several over 100 m in length, temples reaching 11 m high, and platforms some 25 m long were mapped on ridges that flanked an arroyo more than one km from the core of El Pilar.

These professionals were able to record three major monumental architectural features among the 150 residential units. These monuments are within 2 km of the El Pilar core and comprised the minor centers OxTeXik east of the arroyo, Amatal west of the arroyo, and Kum to the north on the way to Laguna Yaloch. In the course of their field validation over the past four years, and have learned much about the relationship between the landscape and the LiDAR-based visualizations, and importantly between the visualizations and the validation of cultural features.  It really all boils down to putting “boots on the ground.”

Dr. Anabel Ford and her team are now involved in the long process of learning how the potential of LiDAR can be fully integrated into our archaeological tool kit.  As an asset, LiDAR strengthens their ability to understand the landscape, hypothesize ancient land use, and appreciate the variability of features uncovered by this new laser technology, both to archaeology in general and the Maya in particular. At the same time, they are learning new aspects about archaeological discovery: they can now see the geography of the forest floor and variations of the forest canopy.  The LiDAR coverage allows them to recognize the geography, the ridges and lowlands as well as the hills and drainage channels. At least 95% of the laser returns reflect above ground biomass, providing direct information on the current state of the forest, its varied heights and relationship to the topography. The forest environmental variability is the key to interpreting ancient Maya settlement, and the density of settlement is the key to understanding the extent of ancient civic centers.

The study of the potential of LiDAR for archaeology in the tropics will be a long process that requires working with massive amounts of data. On average, there are about 25 laser returns per m2, or 25,000,000 returns per 1 km2! For the 1200 km2 of the recently reported coverage in Guatemala, the numbers are staggering! Only 5% of the returns, that is 1,250,000 per km2, are ground point returns; even that number presents a substantial amount of data.  This is a starting point before any ground validation is initiated using GPS.  Ground truth investigation is not simply putting the destination points into the GPS and trekking out to points of interest. To adequately control the data, fundamental data collection protocols must be clearly established.  With every field site visited, the GPS records the tracks, and for every validated destination, waypoints are generated. Without a protocol these data quickly become unwieldy. Immense amounts of data are generated from such projects and care must be taken to process these data to ensure their utility.

For El Pilar, the team have developed and honed a protocol for field validating the “GoTo” points generated in the lab, where points are extracted from LiDAR images in their project GIS.  Based on the Bone mapping visualization strategy developed with Tom Pingel, they have identified elements that appear as cultural features on the landscape.  Over the past four years they have covered 10 km2 and validated 1,214 “GoTo” points, mapped 1,335 new cultural features, verified 611 domestic structures, and mapped 7 civic monumental sites. They found that 16% of the features were not cultural but natural elements: buttresses of the large trees, frond and seed debris from Cohune palms, and problems with LiDAR data interpretations that they call “pox.”  The international team has tracked findings with a strict protocol of field-to-lab management. A data dictionary is maintained for reference and reconfirmation and all data are archived for future use in the context of the GIS.

“We have just launched the Maya Forest Atlas (http://marc-ucsb.opendata.arcgis.com/) and are working to complete the validation for the remaining 10 km2 of the El Pilar Archaeological Reserve for Maya Flora and Fauna,” writes Dr. Anabel Ford and her joint team.

“While LiDAR is no magic wand, our survey efforts have become vastly more informed and effective by incorporating this technology.  We anticipate that the El Pilar data set on the field validation of the LiDAR features will provide a basis for projecting and interpreting the massive coverage we now have for the Maya area.”