Next Article in Journal
Evapotranspiration Acquired with Remote Sensing Thermal-Based Algorithms: A State-of-the-Art Review
Previous Article in Journal
SAR Tomography Based on Atomic Norm Minimization in Urban Areas
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Remote Sensing
Volume 14
Issue 14
10.3390/rs14143432
remotesensing-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Lidar-Based Aboveground Biomass Estimations for the Maya Archaeological Site of Yaxnohcah, Campeche, Mexico

by
Mariana Vázquez-Alonso
1,*,
David L. Lentz
1,
Nicholas P. Dunning
2,
Christopher Carr
2,
Armando Anaya Hernández
3 and
Kathryn Reese-Taylor
4
1
Department of Biological Sciences, University of Cincinnati, Cincinnati, OH 45220, USA
2
Department of Geography and GIS, University of Cincinnati, Cincinnati, OH 45220, USA
3
Laboratorio de Geomática, Centro de Estudios de Desarrollo Sustentable y Aprovechamiento de la Vida Silvestre (CEDESU), Universidad Autónoma de Campeche, Campeche 24079, Mexico
4
Department of Anthropology, University of Calgary, Calgary, AB T2N 1N4, Canada
*
Author to whom correspondence should be addressed.
Remote Sens. 2022, 14(14), 3432; https://doi.org/10.3390/rs14143432
Submission received: 6 June 2022 / Revised: 6 July 2022 / Accepted: 15 July 2022 / Published: 17 July 2022
Browse Figures
Versions Notes

Abstract

:
This study aims to provide a technique applied to archaeology to estimate lidar-based aboveground biomass (AGB) in contemporary tropical forests surrounding archaeological sites. Accurate AGB estimations are important to serve as a baseline to evaluate the wood resources that the ancient Maya could have used for the development of their cities. A lidar processing model is proposed to study the contemporary forest surrounding the Yaxnohcah archaeological site. As tropical forests are highly diverse environments where species are not uniformly distributed, it was necessary to consider the variation within the forest to obtain accurate AGB. Four vegetation communities were defined from a supervised classification of a Sentinel-2 satellite image. A stratified sample was then selected for the field survey that comprised 73 transects of 500 m2 each. To estimate the transect AGB, we used an allometric equation that requires diameter, height, and wood density measurements for identified species. Linear-derived models provided the relationship between field data with lidar statistics for each vegetation type. Predicted average AGB values agreed with those obtained in the field. However, they significantly differed between vegetation types, averaging 83 Mg/ha for lowland forest, 178 for transition forest, and 215 for upland forest communities. From those results, we created a map with wall-to-wall AGB estimates following the distribution of vegetation classes that could complement archaeological research of past land use. Vegetation classification also helped determine that there is a spatial relationship between vegetation communities and the distribution of archaeological settlement features for the ancient city of Yaxnohcah.

Graphical Abstract

1. Introduction

Aboveground biomass (AGB) can be estimated and mapped over large areas using remote sensing tools. It is important to consider forest structure when defining a procedure for estimating AGB; however, other considerations include the selection of data, proper algorithms, and the potential increase in accuracy if data sources are integrated [1]. Remote sensing estimations are usually based on AGB calculated from forest inventories. Field measurements of trees are taken inside plots, where tree diameter is essential and can be complemented with tree height and species wood density to form allometric equations [2]. Remote sensing data from active (radar) and passive sensors (optical) had been previously used to estimate AGB. Optical data from satellite images can be divided into multispectral and hyperspectral based on the number of bands they contain. Previous studies have shown that optical data such as Landsat 8 and Sentinel-2 data are important information sources for assessing forest AGB [3,4]. The algorithms used to establish AGB estimation models are parametric or nonparametric. Parametric methods are statistical regression methods such as linear regression (LR). Nonparametric methods are Machine Learning (ML) algorithms [4].
Besides the traditional use of aerial and satellite images, lidar (light detection and ranging) imaging technology produces time-efficient and wall-to-wall estimates [5]. Foresters initially used lidar to assess AGB and carbon and to monitor trees and forests at a fine-grained level that previously could only be done with forest surveys over a limited area [6]. Lidar sensors can accurately quantify AGB in a dense tropical forest [5,7]. However, they do not provide forest species information. For this reason, it is critical to unite lidar data with multispectral satellite image analysis and forest surveys. AGB biases using allometric equations can be evaluated if lidar-based AGB estimates and on-the-ground tree identification and diversity measurements are made [8,9].
The use of lidar in archaeology has greatly simplified the identification of anthropogenic Precolumbian land-use and settlement patterns [10,11]. Recent studies with lidar have uncovered a landscape in the Maya Lowlands more modified than expected for intensive agriculture. After this finding, Canuto et al. [12] proposed using lidar data to obtain AGB quantification of the current forest, which can deepen the understanding of its relationship with ancient environments. A study at the Maya site of Tikal presented a land use model that reflects the interaction between the ancient Maya and their tropical forest. The evidence was recovered from various data sources, including paleoethnobotanical and ecological as AGB estimations from field surveys in the modern forest [13,14].
The ancient city of Yaxnohcah, located in southern Campeche, México, is situated between two large karst depressions, locally known as bajos, surrounded by rolling karst uplands [15]. This landscape has shaped the development of the vegetation communities. Yaxnohcah arose as a major Maya urban center in the Preclassic period (ca. 1000 BCE–250 CE) [16]. While the construction of large architectural complexes waned in the Classic period (ca. 250–850 CE), Yaxnohcah remained a significant population center until fading and ultimate abandonment during the Postclassic period (1000–1400 CE). The 2000 years of Maya occupation significantly altered the landscape within and around Yaxnohcah, including the construction of numerous limestone and plaster structures, the erosion of soil from sloping lands and deposition of sediments across adjacent lowlands, and partial removal of forest for agriculture and management of remaining woodlands [15]. After centuries of abandonment, the tropical forest that reoccupied this altered landscape would not have been identical to that first encountered by the initial Maya Preclassic colonists. However, an archaeobotanical study at Tikal, Guatemala, found a co-occurrence of forest tree species comparing preserved wood from ancient contexts with the modern forest [17].
Archaeological studies have benefited from scanning their sites with lidar sensors, especially the Maya area covered with dense tropical forests. However, the lidar point cloud data have not been completely explored for forest research related to the archaeological sites at the landscape level. To have not only a broader view of the landscape, but also to understand the specific use of the vegetation in ancient times, we aim to unite the information provided by landscape ecology and forest surveys. Our main objective was to estimate above ground forest AGB in the contemporary Calakmul forest and identify areas disturbed by the ancient Maya occupation of Yaxnohcah, Campeche, Mexico, using airborne lidar data. In doing so, we needed to (1) delineate the distribution of the contemporary forest communities that envelope the landscape surrounding the archaeological site of Yaxnohcah, (2) study the relationship between lidar-derived forest structure metrics and AGB in the different vegetation communities of the forest, and (3) determine if there is a spatial relationship between vegetation types and archaeological settlement patterns.

2. Materials and Methods

2.1. Survey Area and Landscape

Our survey area is in the state of Campeche in the south of Mexico, near the border of Guatemala and Belize (Figure 1). The area has an elevation between 230 and 285 masl. The climate is tropical sub-humid with mean annual rainfall between 1200 and 1500 mm, most of which falls from June to October, while during the dry season from November to May it rains less than a third of the total. During the rainy season, a series of natural water reservoirs, or possibly human-made, known in the region as aguadas, are filled. Most of the water in these deposits dries up by the end of the dry season. Overall, the area lacks perennial surface water. The average annual temperature is 27 °C with average extremes of 20 and 34 °C [18]. It is located within the Calakmul Biosphere Reserve, which includes 723,000 ha of tropical forest. The Calakmul forest was established as a Biosphere Reserve in 1989.
Holdridge’s [19] life zones situate the Calakmul region under the moist semitropical forest zone. A similar vegetation type was assigned by [20], high or medium semi-evergreen forest. The classification proposed by [21] has been mainly used to recognize the vegetation and distribution of the area. A floristic list for the Calakmul region found 1537 species after extensive field survey [22]. Based on it, Martínez and Galindo-Leal [23] undertook a detailed classification of the Calakmul forest that describes vegetation associations characterized by the relative dominance of species.

2.2. Vegetation Classification

To understand the archaeological site of Yaxnohcah at the landscape level, we defined vegetation communities in the modern forest using ground surveys coupled with lidar data and multispectral satellite imagery in a 25 km2 area. Lidar data was collected on 23–24 May 2014, with the sensor measuring at a density of 15 pulses/m2 [24]. For our initial analysis, we classified the vegetation using an Isodata clustering unsupervised classification from a Landsat 8 satellite image collected on 25 February 2015, to select ground survey locations [25]. The preliminary research, joined with ground surveys, allowed us to have validation points to perform a Random Forest supervised classification of a Sentinel-2 satellite image collected on 24 February 2017, from Earth Explorer (https://earthexplorer.usgs.gov/ accessed on 1 October 2021). All bands except atmospheric were converted to 10 m resolution using the Sentinel Application Platform (SNAP) [26]. We compared the areas obtained from the classification with tree height breaks obtained from the lidar’s digital surface model (DSM) and digital elevation model (DEM) difference for more accurate interpretations. We defined four vegetation classes: upland forest (high semi-evergreen forest), transitional forest (medium semi-evergreen forest), lowland forest (low semi-evergreen forest), and wetland (aquatic plants). Lowland forests occur in seasonal swamps locally known as bajos, whereas aguadas refer to the small perennial wetlands and ponds. Finally, a map of vegetation communities was created, and their areas were calculated using the Sentinel-2 Toolbox [26].

2.3. Forest Survey Transects

A forest survey was conducted between 2017 and 2019 in the 25 km2 area covered by the lidar flights. After defining the forest vegetation communities as upland forest, transitional forest, lowland forest, and wetland, we carried out quantitative surveys of the four vegetation types in transects of 500 m2 each (10 × 50 m). For all stems (tree, shrub, or vine) greater than 6 cm diameter (D) at 130 cm aboveground, height (H) in m, XY locations within the transect and species name were recorded. In addition, a sub-quadrant of 25 m2 was sampled for the low vegetation forest (bajo), measuring stems between 1 and 6 cm in diameter. The forest survey was completed with the help of local people, who provided common names and economic uses of species when known. Their knowledge of the forest was essential for the accuracy and completion of the survey. The list of the tree species found during the forest survey is added as Supplementary Material.
AGB was calculated for each stem (>6 cm D) in the transects to obtain the wood availability from the forest. The allometric equation AGB est = 0.0673 × (ρD2H)0.976 [27] specific for tropical vegetation was used because it also includes the tree species identification. Where ρ is wood density (g cm−3), obtained from the DRYAD world database and additional wood density data [28], D is stem diameter (cm), and H is total tree height (m); results are given in megagrams (Mg) [27]. To calculate carbon stocks (or aboveground carbon AGC), a conversion was made from the mean AGB to a carbon ratio of 0.471 [29]. All statistical analyses were performed using the BIOMASS R package of the R statistical software [30,31]. Additionally, basal area (Basal area m2 = pi × D2/40000), a measure of tree density was calculated, which is known to be correlated with AGB [32,33].

2.4. Lidar Aboveground Biomass Estimation

2.4.1. Lidar Data Processing

The lidar point cloud already classified for ground points by The National Center for Airborne Laser Mapping NCALM, was verified in 3D and transect views with Global Mapper pro GIS software [34] (Figure 2). Lidar data tiles have a return density per square meter of 1–102 with an average of 20 points. A reference image with a 1 m pixel size was created using the returns’ intensity with the FUSION package [35].
Lidar returns that occurred within each of the 73 transects (21 lowland, 27 transition, 24 upland, and 1 wetland) were extracted to create a point cloud for each transect. A digital elevation model, DEM, was created using preassigned ground points, in which the raster pixels represent bare-earth elevations above sea level. For normalizing, the ground surface elevation obtained from the DEM was subtracted from all returns to convert from altitude to height. After normalizing the lidar data, a digital surface model, DSM, and a canopy height model, CHM, raster layers with 1 m resolution were created using FUSION. The workflow of the satellite and lidar data processing used in this study is shown in Figure 3.

2.4.2. Linear Models

Linear models were used to define relationships at the transect level, between lidar metrics obtained from the point cloud and AGB values estimated from field measurements. Descriptive statistics of the lidar data were computed for each transect, using all returns above 1 m height to avoid the herbaceous layer (Appendix A). A canopy overstory threshold height of 3 m was used to compute lidar canopy cover metrics. Lidar metrics from the transect areas were summarized for regression modeling and were evaluated for normality with Shapiro–Wilk tests.
Lidar-derived predictor variables were selected using the Rcmdr package [36] in R v.4.1.0 [30]. Pearson’s correlation coefficient was used to measure the linear relationship between variables, the coefficients were represented in a histogram matrix. A linear regression formula was constructed for each vegetation type, using variables that indicate a strong correlation between the transect lidar metrics and the AGB measured in the field.

2.4.3. GIS Mapping

To predict AGB, we used the lidar-derived forest structure metrics selected for the models for the entire study area covered by lidar (25 km2). Lidar statistics related to tree heights based on the CHM were obtained with FUSION software for a 22.4 m grid that corresponded to the 500 m2 sampling plots; data were then assigned to grid cells created in a QGIS layer. Linear models obtained for the classified vegetation were the base to extrapolate the AGB calculation. Grid AGB values were obtained from multivariate linear models and could be mapped with QGIS [37].

2.5. Forest Area Affected by Settlements

The extent of the ancient city of Yaxnohcah was estimated to be within the 25 km2 sampled by the lidar and forest surveys; however, the ancient limits of the city are unknown. The distribution of settlements in the ancient city has been studied from 2011 until the present by the Yaxnohcah Archaeological Project (YAP) [24,38,39]. The project provided data on previously recognized and ground-truthed settlements from the lidar DEM. The area occupied by the recognized settlement features was compared with the vegetation types delimited from the modern forest. A map of the areas impacted by settlement was created, and areas were quantified. The study area was also divided into five topographic zones: bajo, bajo margins, upper terraces (mesoland), uplands and highlands, where the archaeological mounds were located to obtain a density ratio.

3. Results

3.1. Landscape Vegetation Communities

Ground surveys were carried out at four classes of assigned vegetation: upland forest, transition forest, low forest (bajo), and localized wetlands (aguadas). The distribution of vegetation types in the landscape relates to changes in topography, soil composition, and seasonal duration of standing water. After verifying ecotones in each vegetation type, a Random Forest supervised classification of a Sentinel-2 satellite image was completed. Tree heights obtained from the lidar’s CHM were combined with the classification for more accurate interpretations of the vegetation (Figure 4).
Based on the Sentinel-2 image classified, the current vegetation surrounding the Yaxnohcah site is wetlands 0.025%, lowland forest 21.8%, upland forest 36.1%, and transitional forest 41.9%. We found ecotones in the vegetation where one species is more abundant. As part of the lowland vegetation, there is an area covered only by palo de tinto Haematoxylum spp. with an open understory, while another lowland had an understory of sedge (Cyperaceae). Additionally, we found Pachira aquatica Aubl. (bobal) trees in an aguada, a unique feature in the area.

3.2. Forest Composition

During fieldwork, we sampled 73 plots of 500 m2 each, 3.65 ha in total, from the 25 km2 of the study area. We measured 4789 individuals with a diameter greater than 6 cm, and 94% were identified as species. We found 163 morphospecies, including 136 identified at a species level, eight at the genus level, and five at the family level, all distributed across 37 families; 14 morphospecies were not identified. The heights of the trees measured in the transects varied from 1.5 to 40 m, with a mean of 9.2 m. Heights were on average 6.4 m (±2.4) for bajo, 10.5 m (±4.5) for transition, and 12.1 m (±5.2) for upland. A one-way ANOVA comparing the heights of the three vegetation types revealed that there was a statistically significant difference between at least two groups (F (24,786) = 904.5, p < 0.01). Tukey’s HSD test for multiple comparisons found that the mean value of the three groups significantly varied (p < 0.0001 at 95%).
Heights obtained from lidar were on average 10.2 m, with a maximum of 42.4 m. Heights were on average 6.5 m (±2.1) for bajo, 10.9 m (±1.9) for transition, and 12.0 m (±2.3) for upland. From the collected data, it is possible to estimate the total number of trees in the area for each species. We used the abundance found in the transects and extrapolated it for the area covered by each type of vegetation. For a single species, the total abundance was as high as 30,000 trees per square kilometer (Appendix B).

3.3. Aboveground Biomass Estimation

3.3.1. Forest Survey AGB Estimation

The objective was to estimate AGB for the Yaxnohcah forest considering the different vegetation types. We used two approaches, field data and metrics obtained from lidar. During fieldwork, we measured 4789 individuals, including trees, shrubs, and vines above 6 cm in diameter. The mean diameter of all trees measured was 12.8 cm and ranged up to 126 cm, and the basal area average was 25.08 m2 ha−1.
The allometric equation selected AGB est = 0.0673 × (ρD2H)0.976 [23] is based on diameter, height, and wood density, and gave us a general value of 179.01 Mg/ha for field data. However, there is variability in forest AGB within the three vegetation types (Figure 5). AGB ranged from 80 Mg/ha in low forest to 197 Mg/ha in transitional forest and 241 Mg/ha in upland forest (Table 1). For the lowland we additionally measured trees in 13 transects with a <6 cm D in plots of 25 m2 at the beginning of the main transects. The small trees add an average of 14.5 Mg/ha to the AGB. In the total 3.65 ha surveyed, the species with higher AGB Mg/ha were Manilkara zapota (L.) P.Royen (34.9), Brosimum alicastrum Sw. (24.6), Pouteria reticulata (Engl.) Eyma (19.41), and Bucida buceras L. (10.69). To obtain Aboveground Carbon (AGC), we used 0.471 as AGB’s ratio, obtaining 38 Mg/ha in lowland forest, 93 Mg/ha in transition forest, and 113 Mg/ha in upland forest.

3.3.2. Lidar AGB Estimation

Lidar metrics were obtained for the complete study area using the FUSION program. Data from the 57 metrics were assigned to a grid created in QGIS, which corresponds to the transect sizes. Using the Sentinel vegetation classification, this variable was also assigned to each grid cell.
Multivariate linear models were developed for each vegetation type using R (Table 2), where AGB is the response variable to lidar metric variables from each transect. Variables were transformed when necessary to normalize the data. Finally, the linear models were applied to each cell in the grid for the corresponding vegetation type.
For the lowland forest, the more significantly correlated metrics were the Median of the Absolute Deviations from the overall mode (MAD mode) and Average Absolute Deviation (AAD). These metrics are in the lower height quartiles, opposite to transition and upland, where the significant metrics were more in the higher quartiles. For transition, the metrics were percentile value 95th (P95), and All returns Above Mode (AAM), and for upland percentiles values 80th and 90th (P80 and P90). Predicted AGB values from lidar models for each transect (Appendix C) show similar means.
AGB models were extrapolated to the complete landscape at the transect resolution per vegetation type (Figure 6). The predicted AGB values are on average, 83 Mg/ha for lowland, 178 for transition, and 215 for upland. The total AGB available for the area is 415,000 Mg, with each vegetation type supplying a different proportion 13% lowland, 43% transition, and 44% upland.

3.4. Forest Area Affected by Settlements

We used the settlement location provided from archaeological survey records to compare with the vegetation distribution in the landscape. We determined that settlements spread over an area of 2410.5 ha and cover an area of 50.5 ha. According to the vegetation distribution map, they are located to a greater extent in the current upland forest (62.1%) rather than in transitional forest (26.7%) or lowland forest (10.9%) (Figure 7). The settlements located in lowland vegetation are in smaller bajos (also known as “pocket bajos”), rather than in larger bajos that are seasonally inundated areas. Due to the different criteria of subdivisions of the topographic zones, the results are slightly different from the vegetation allocation (Table 3). Although, they concur in allocating the highest percentage of mound frequency in the mesoland and upland topographic zones (65.6%) (Figure 8).

4. Discussion

Our initial aim was to estimate the potential forest resources that the ancient Maya had to establish and maintain the city of Yaxnohcah. To do this, we used AGB measurements in the contemporary forest and identified areas selected for ancient Maya occupation to compare with the present vegetation distribution.
Semi-evergreen tropical forests cover the present landscape surrounding the archaeological site of Yaxnohcah. It was important for us to acknowledge that the forest is not uniform in composition, and that the vegetation types could have offered different resources to the Maya population. Because of this, we performed a vegetation classification and a stratified sampling in which the forest was assigned to four main types of vegetation: upland, transitional, low, and wetland. Wetland areas were excluded from the quantifications as they are mostly composed of herbaceous vegetation.
Upland vegetation trees in the transects had an average height of 12.1 m (±5.2), giving the highest AGB with 241 Mg/ha and are found in 36.1% of the area. The AGB for upland forest at Yaxnohcah is higher than what Read and Lawrence [40] found in their study at Arroyo Negro, located within the Calakmul municipality but outside the reserve. For mature forests, they estimated 125.94 (±12.6) Mg/ha for an area that was selectively logged in the past half-century but not cultivated, as is the case of the Yaxnohcah forest. Their estimated wood specific gravity at sampling locations was 0.70 g/cm3, and our average wood density was 0.67 g/cm3. Some species with high wood density at Yaxnohcah are Krugiodendron ferreum (Vahl) Urb. (1.35 g/cm3), Bucida buceras (1.03 g/cm3), and Manilkara zapota (0.99 g/cm3). Cairns [41] estimated 225 Mg/ha for a dry semi-evergreen intact forest in southern Quintana Roo state. That forest presents a similar composition to Yaxnohcah; however, the maximum height is 30 m, while the maximum height we recorded is 40 m.
Transitional forest is the most abundant vegetation class, located across 41.9% of the study area; its trees have an average height of 10.5 m (±4.5), and it has 197 Mg/ha of AGB. Low forest is less abundant, covering 21.8% of the area, and whose trees are shorter with an average of 6.4 m (±2.4) and produce 80 Mg/ha of AGB. In comparison, a forested area near Xpujil, located 77 km northeast of Yaxnohcah and drier, has AGB estimations of 35.4–72.7 Mg ha−1 for semi-perennial medium forest and AGB values of 77.5–94.9 Mg ha−1 for inundated forests [38].
Assuming that AGB accumulation follows a linear trend, the regression analysis AGB (Mg ha−1) = 11.431 + 2.615 × age (R2 = 0.57, n = 28, p < 0.0001) indicates that secondary forest AGB is related to age [40]. Using this regression line, the estimated minimum age for the Yaxnohcah forest, lowland is 26 years, the transitional is 72 years, and the upland is 87 years. Their estimate for forest recovery to a pre-logged state is 65–120 years. Even though this regression seems inaccurate for the low forest, the analysis indicates that the forest at Yaxnohcah is a mature forest.
AGB studies over broader areas of the Yucatan Peninsula have employed different technologies such as Goddard’s Lidar Hyperspectral and Thermal Imager (G-LiHT) or ALOS PALSAR backscatter data, trained with existing field inventory data [42,43]. For the Yaxnohcah study, it was necessary to produce AGB models for each vegetation type due to the equations including tree heights, and this measurement is significantly different between vegetation types. Lidar models created did not respond equally for each vegetation type. The model for low vegetation model explained the 59% of the variation in the regression, even though this vegetation has more uniformity in lidar tree heights (6.48 m average, ±2.11). One reason for this variation could be because we have fewer transects in this vegetation (n = 21); another possibility is that the AGB equation is more suitable for taller trees. However, in comparing the performance of specific AGB equations generated for small-sized trees (<10 cm D) for available species [41,44], there is not a significant difference with the general equation used in our study (t (8) = −0.891, p = 0.398).
The transitional forest linear model explained just 27% of the AGB variation. The variability of heights and diameter in this vegetation class did not offer a trend that lidar metrics could explain or that the allometric equation calculated accurately. In part, the higher basal area in the trees was not always related to taller trees but gave a higher AGB. We surveyed more area for this vegetation (27 transects) than for the other two vegetation types because of the variability we found. A possibility for further study of this vegetation is to acquire more lidar points per square meter that would better represent the density of this forest.
Upland forest models had the best fit, 73% of the AGB is explained with the percentile values 80th and 90th of the lidar height. Metrics of the upper level lidar heights are easier to model against tree heights, and as height is a main part of the AGB equation, it offers a precise comparison. From the field survey, we measured maximum tree heights of 40 m, and from the lidar data, the maximum height was 42 m. More tall trees were located in the upland forest than elsewhere.
Local uses of trees in the present-day give us an idea of wood characteristics that could make them suitable for similar use in the past. The Maya could obtain species for timber, such as Aspidosperma desmanthum Benth. ex Müll.Arg. from upland vegetation, Vitex gaumeri Greenm. (yaxnik) from transitional forest, or Metopium brownei (Jacq.) (chechem) Urb. from lowland forest (bajos). Some of these species are also sources of edible fruits including Cordia dodecandra A.DC. (ciricote), or Croton arboreus Millsp. (cascarillo) for medicine. While species from bajos, such as Hyperbaena winzerlingii Standl. (naranjillo), are preferred for firewood, because even when they have a smaller trunk diameter, some also have high wood density. Bajo species generally develop smaller in size and have a fast regrowth, this makes them more suitable for firewood harvesting and transport. Specifically, the smaller size of the trees in lowland forest could have made the harvesting easier for women and young people who likely collected fuel on a daily basis.
The current forest composition at the Calakmul Biosphere Reserve is known from floristic studies [22,23]. Additionally, historical descriptions of the Calakmul vegetation refer to some species that were found during previous times [45]. However, the connection of modern composition with Maya times comes from the archaeological record. Identifications of pollen, charcoal, and eDNA recovered from excavations at Yaxnohcah show some dominant species present in the modern forest, such as Brosimum and Manilkara as well as Pinus that is not there today [46]. Supported by the identification of archaeobotanical remains, it is confirmed that the forest composition in ancient Maya times around Yaxnohcah had similarities to that of the present-day, though, as noted in the introduction, acknowledging that the long Maya occupation itself changed the nature of the landscape and forest. Botanical remains reflect ancient forest composition around the archaeological settlement [47]; however, they represent individual trees, not the complete vegetation composition or their abundance. Because of this, the modern forest composition and species abundance can be useful for interpretations of resource availability in the past.
The present-day landscape is covered by semi-evergreen tropical forests lying inside a biosphere reserve that has been preserved for the last 30 years. Before that time, selective logging of valuable timber species such as Swietenia macrophylla King (mahogany, caoba) and Cedrela odorota L. (spanish cedar, cedro) occurred. Because of logging, the forest is missing some dominant trees of the upland vegetation that can reach 35 m tall and 1.7 m in diameter [16] and could have a density of 6 trees/ha. Chicle sap extraction from Manilkara zapota trees was a common practice. However, its use does not seem to have diminished the forested area but could have modified its composition due to chicle trees being selectively protected at the expense of other species. Palo de tinto (Haematoxylum campechianum L.) extraction was not reported from southern Campeche, compared to how important it was for trading in the north of the Yucatan Peninsula [48]. Overall, the vegetation suffered from extraction but has never been cleared for cultivation in the modern times. When the Maya settled in, what is now southeastern Campeche during the Middle Preclassic period, they arrived at an unoccupied area. Upland forest was likely the most affected during Maya times by settlements, with 62% of Yaxnohcah structures located within this vegetation class. The Maya inhabitants probably had several reasons for selecting these locations, including higher elevations that were less prone to flooding, as well as selecting for the proximity of desirable forest species. As far as we know, this forest has never been more modified than it was in Maya times, when they were practicing an active management of the landscape for agriculture and forestry.
A palaeoecological study at the ancient Maya city of Tikal in Guatemala estimated higher AGB values [13,14] than the AGB estimated for Yaxnohcah. From the relationship between the forest inventory and the Landsat satellite image bands, they calculated 289 Mg/ha (±26) for upland forests and 182 Mg/ha (±5.23) for bajo areas. The AGB for upland forest at Tikal is 17% higher than at Yaxnohcah; this increase can be related to the Yucatan Peninsula’s precipitation gradient that increases NW towards the SE [49]. To evaluate if the wood resources estimated for Yaxnohcah were sufficient, it is necessary to take into account population size and settlement distribution. This evaluation could give us an idea if fewer resources supported less people than Maya cities with higher AGB forest as Tikal that were more extensive and supported more people, AGB being a limiting factor for population size. Further interpretations of the AGB resources at Yaxnohcah are necessary to understand their use for the development of the city.

5. Conclusions

The aim of this study was to accurately estimate the AGB of the contemporary forest surrounding the Yaxnohcah archaeological site. This calculation provided baseline data that enabled us to assess the wood resources that the ancient Maya potentially had available for the city development. We used linear models to estimate relationships between lidar metrics and data obtained from a stratified forest survey conducted by measuring trees encountered in 500 m2 transects. AGB was estimated for the entire area following vegetation classes defined from a satellite image supervised classification. We noticed that the lidar tree height averages were quite similar to the field measurements of trees inside the transects and provided a pattern of tree heights that compares to the distribution of vegetation communities. Although the contribution of height in allometric equations is important for AGB estimation, the contribution of diameter is equally essential, especially for smaller or shorter trees (low forest/transitional forest). To find better relationships between the forest and lidar data, it may be worth developing allometric equations for each vegetation community due to the different growth patterns of fully grown trees.
The estimated AGB of upland forest is 22% higher than for transitional forest and 300% higher than for the low or bajo forest. On one side, the substantial difference in AGB between vegetation communities shows the necessity of considering the differences in vegetation and conducting stratified sampling. On the other side, AGB calculations using lidar metrics seem to be mostly related to the tree height, for which field sampling could be based on the difference between heights provided by the same lidar point cloud.
Obtaining wall-to-wall AGB estimates in the area where an archaeological site is located is valuable for interpretations of ancient land use. In this case, the modern upland forest distribution coincides with more presence of settlements of the Yaxnohcah ancient city. To what extent the forest was affected by the ancient Maya needs support from palaeoecological studies, including pollen, charcoal, and DNA, to be added to AGB studies to aid in past land use interpretation [42].
Wood resources were estimated from AGB values at the landscape level and by vegetation communities. In addition, the abundance and volumes of useful tree species were determined. In the case of firewood species defined from charcoal found in an archaeological context, these results could be associated with a vegetation class from which they were extracted and with an abundance estimated from modern forest counts. Archaeological information, such as settlement distribution or population density, analyzed together with the AGB, can help bring about a greater understanding of forest management practices of ancient cultures. The use of lidar in archaeological studies has advanced the field in tropical areas [50], particularly for the ancient Maya culture [12,51]. Improvements in lidar technology, such as increased points per square meter, will benefit archaeological interpretations of the terrain as well as the canopy surface. Because of this, archaeological projects would be well-served to consider using lidar data for environmental studies as vegetation surveys in conjunction with their excavation plans.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/rs14143432/s1, Table S1: List of plant species found in the Yaxnohcah forest survey (2017, 2018, and 2019).

Author Contributions

Conceptualization, M.V.-A., D.L.L. and N.P.D.; methodology, M.V.-A., D.L.L. and C.C.; software, M.V.-A., C.C. and A.A.H.; validation, M.V.-A. and C.C.; formal analysis, M.V.-A.; investigation, M.V.-A.; resources, K.R.-T., D.L.L., N.P.D. and A.A.H.; data curation, M.V.-A.; writing—original draft preparation, M.V.-A.; writing—review and editing, D.L.L., N.P.D., C.C. and A.A.H.; visualization, M.V.-A.; supervision, D.L.L.; project administration, K.R.-T., D.L.L., N.P.D. and A.A.H.; funding acquisition, D.L.L., N.P.D., K.R.-T. and M.V.-A. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the United States National Science Foundation [Grants BCS-1632392 (ND and DL)], the Social Sciences and Humanities Research Council of Canada Grants 430-2017-00190 and 892-2019-3070 (KR-T), a University of Calgary URGC Seed Grant (KR-T), the Universidad Autónoma de Campeche (AA), the Graduate School Dean’s Dissertation Completion Fellowship UC (MV), and Consejo Nacional de Ciencia y Tecnología (MV).

Data Availability Statement

The data presented in this study are available in the article.

Acknowledgments

We would like to thank the local people from the Calakmul area and Hannah Clements for their support during fieldwork.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Summary of the lidar forest structure variables using elevation. The statistics were derived from the lidar point cloud for each ground transect and for each 500 m2 grid cell in the 25 km2 study area using FUSION [35]. The site was stratified in three vegetation types, upland forest, transition forest, and lowland forest. The height aboveground of each return was obtained by subtracting the DEM from the DSM prior to computing these metrics.
Elev minimum (minimum height above ground)
Elev maximum (maximum height above ground)
Elev mean (mean height above ground)
Elev mode (mode height above ground)
Elev stddev (standard deviation of height above ground)
Elev variance (variance of height above ground)
Elev CV (coefficient of variation of height above ground)
Elev IQ (interquartile distance)
Elev skewness (skewness of height above ground)
Elev kurtosis (height kurtosis of height above ground)
Elev AAD (average absolute deviation from mean height)
MADMedian (median of the absolute deviations from the overall median)
MADMode (median of the absolute deviations from the overall mode)
L-moments (L1, L2, L3, L4)
L-moment skewness
L-moment kurtosis
Percentile values (1st, 5th, 10th, 20th, 25th, 30th, 40th, 50th, 60th, 70th, 75th, 80th, 90th, 95th, 99th percentiles)
Canopy relief ratio ((mean − min)/(max – min))
Generalized means for the 2nd and 3rd power (Elev quadratic mean and Elev cubic mean)

Appendix B

Table
Table A1. Estimated number of trees per square kilometer for the most abundant species per vegetation type.
Table A1. Estimated number of trees per square kilometer for the most abundant species per vegetation type.
Low Forest SpeciesTrees/km2Transition Forest SpeciesTrees/km2Upland Forest SpeciesTrees/km2
Eugenia winzerlingii Standl.29,238Pouteria reticulata (Engl.) Eyma31,778Pouteria reticulata25,667
Metopium brownei (Jacq.) Urb.20,190Drypetes lateriflora (Sw.) Krug & Urb.7333Trichilia minutiflora15,583
Haematoxylum campechianum L.13,619Manilkara zapota (L.) P.Royen6667Brosimum alicastrum13,417
Hyperbaena winzerlingii Standl.13,619Lonchocarpus xuul Lundell5630Drypetes lateriflora7083
Cameraria latifolia L.12,476Nectandra coriacea (Sw.) Griseb.5037Melicoccus oliviformis Kunth3833
Haematoxylum calakmulense8381Krugiodendron ferreum (Vahl) Urb. 3111Manilkara zapota2667
Gymnopodium floribundum Rolfe7429Brosimum alicastrum Sw.3037Protium copal (Schltdl. & Cham.) Engl.1583
Croton arboreus Millsp.7333Gymnopodium floribundum3037Mosannona depressa (Baill.) Chatrou1583
Erythroxylum areolatum L.6952Lonchocarpus guatemalensis Benth.2519Lonchocarpus xuul1167
Eugenia aeruginea DC.5619Trichilia minutiflora Standl.2370Dendropanax arboreus (L.) Decne. & Planch.1167

Appendix C

Table
Table A2. Aboveground biomass estimates calculated by fieldwork survey and lidar predicted for each transect.
Table A2. Aboveground biomass estimates calculated by fieldwork survey and lidar predicted for each transect.
Low VegetationTransition VegetationUpland Vegetation
TransectField Mg ha−1Lidar Mg ha−1TransectField Mg ha−1Lidar Mg ha−1TransectField Mg ha−1Lidar Mg ha−1
183.333.37018.038.970522.8821.90
193.493.93028.288.000613.5913.54
205.925.430316.418.370722.1317.43
213.914.600419.5314.09087.837.72
315.074.761114.6711.45099.509.12
324.563.691215.0811.981011.839.88
332.913.69139.8412.292220.2811.70
345.674.07147.7212.10237.3112.35
351.051.791511.9210.852412.999.94
364.003.53165.737.04256.136.29
373.804.72176.418.64263.714.95
384.584.562811.056.31274.284.20
393.523.954610.019.08295.716.28
406.445.684710.3410.03304.735.11
414.114.84488.2411.305612.8713.76
423.344.11498.8110.99619.429.63
433.582.905017.6014.58628.6110.76
443.292.765110.8611.50649.848.75
452.563.325211.1410.026511.5511.48
594.153.21535.0611.05669.2514.22
634.785.145410.4411.356710.2414.18
556.068.516814.5212.39
604.906.057512.0616.42
705.307.76
7111.268.35
725.706.90
744.948.70
Mean4.004.00 9.839.86 10.9210.96
Total84.0384.03 265.36266.28 251.26252.00

References

  1. Zhang, W.; Zhao, L.; Li, Y.; Shi, J.; Yan, M.; Ji, Y. Forest Above-Ground Biomass Inversion Using Optical and SAR Images Based on a Multi-Step Feature Optimized Inversion Model. Remote Sens. 2022, 14, 1608. [Google Scholar] [CrossRef]
  2. Gibbs, H.K.; Brown, S.; Niles, J.O.; Foley, J.A. Monitoring and Estimating Tropical Forest Carbon Stocks: Making REDD a Reality. Environ. Res. Lett. 2007, 2, 045023. [Google Scholar] [CrossRef]
  3. Li, Y.; Li, M.; Li, C.; Liu, Z. Forest Aboveground Biomass Estimation Using Landsat 8 and Sentinel-1A Data with Machine Learning Algorithms. Sci. Rep. 2020, 10, 9952. [Google Scholar] [CrossRef]
  4. Moradi, F.; Darvishsefat, A.A.; Pourrahmati, M.R.; Deljouei, A.; Borz, S.A. Estimating Aboveground Biomass in Dense Hyrcanian Forests by the Use of Sentinel-2 Data. Forests 2022, 13, 104. [Google Scholar] [CrossRef]
  5. D’Oliveira, M.V.N.; Broadbent, E.N.; Oliveira, L.C.; Almeida, D.R.A.; Papa, D.A.; Ferreira, M.E.; Zambrano, A.M.A.; Silva, C.A.; Avino, F.S.; Prata, G.A.; et al. Aboveground Biomass Estimation in Amazonian Tropical Forests: A Comparison of Aircraft- and GatorEye UAV-Borne LiDAR Data in the Chico Mendes Extractive Reserve in Acre, Brazil. Remote Sens. 2020, 12, 1754. [Google Scholar] [CrossRef]
  6. Rosette, J.; Surez, J.; Nelson, R.; Los, S.; Cook, B.; North, P. Lidar Remote Sensing for Biomass Assessment. In Remote Sensing of Biomass—Principles and Applications; Fatoyinbo, L., Ed.; InTech: London, UK, 2012; ISBN 978-953-51-0313-4. [Google Scholar]
  7. D’Oliveira, M.V.N.; Reutebuch, S.E.; McGaughey, R.J.; Andersen, H.-E. Estimating Forest Biomass and Identifying Low-Intensity Logging Areas Using Airborne Scanning Lidar in Antimary State Forest, Acre State, Western Brazilian Amazon. Remote Sens. Environ. 2012, 124, 479–491. [Google Scholar] [CrossRef]
  8. Clark, M.L.; Roberts, D.A.; Ewel, J.J.; Clark, D.B. Estimation of Tropical Rain Forest Aboveground Biomass with Small-Footprint Lidar and Hyperspectral Sensors. Remote Sens. Environ. 2011, 115, 2931–2942. [Google Scholar] [CrossRef]
  9. Phillips, O.L.; Sullivan, M.J.P.; Baker, T.R.; Monteagudo Mendoza, A.; Vargas, P.N.; Vásquez, R. Species Matter: Wood Density Influences Tropical Forest Biomass at Multiple Scales. Surv. Geophys. 2019, 40, 913–935. [Google Scholar] [CrossRef] [Green Version]
  10. Garrison, T.G.; Houston, S.; Alcover Firpi, O. Recentering the Rural: Lidar and Articulated Landscapes among the Maya. J. Anthropol. Archaeol. 2019, 53, 133–146. [Google Scholar] [CrossRef]
  11. Golden, C.; Murtha, T.; Cook, B.; Shaffer, D.S.; Schroder, W.; Hermitt, E.J.; Alcover Firpi, O.; Scherer, A.K. Reanalyzing Environmental Lidar Data for Archaeology: Mesoamerican Applications and Implications. J. Archaeol. Sci. Rep. 2016, 9, 293–308. [Google Scholar] [CrossRef] [Green Version]
  12. Canuto, M.A.; Estrada-Belli, F.; Garrison, T.G.; Houston, S.D.; Acuña, M.J.; Kováč, M.; Marken, D.; Nondédéo, P.; Auld-Thomas, L.; Castanet, C.; et al. Ancient Lowland Maya Complexity as Revealed by Airborne Laser Scanning of Northern Guatemala. Science 2018, 361, eaau0137. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  13. Lentz, D.L.; Magee, K.; Weaver, E.; Jones, J.G.; Tankersley, K.B.; Hood, A.; Islebe, G.; Hernandez, C.E.R.; Dunning, N.P. Agroforestry and Agricultural Practices of the Ancient Maya at Tikal. In Tikal: Paleoecology of an Ancient Maya City; Cambridge University Press: Cambridge, UK, 2015; pp. 152–185. [Google Scholar]
  14. Lentz, D.L.; Dunning, N.P.; Scarborough, V.L.; Magee, K.S.; Thompson, K.M.; Weaver, E.; Carr, C.; Terry, R.E.; Islebe, G.; Tankersley, K.B.; et al. Forests, Fields, and the Edge of Sustainability at the Ancient Maya City of Tikal. Proc. Natl. Acad. Sci. USA 2014, 111, 18513–18518. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  15. Dunning, N.P.; Anaya Hernández, A.; Beach, T.; Carr, C.; Griffin, R.; Jones, J.G.; Lentz, D.L.; Luzzadder-Beach, S.; Reese-Taylor, K.; Šprajc, I. Margin for Error: Anthropogenic Geomorphology of Bajo Edges in the Maya Lowlands. Geomorphology 2019, 331, 127–145. [Google Scholar] [CrossRef]
  16. Reese-Taylor, K. Founding Landscapes in the Central Karstic Uplands. In Maya E Groups: Calendars, Astronomy, and Urbanism in the Early Lowlands; University Press of Florida: Gainesville, FL, USA, 2017; pp. 480–514. ISBN 978-0-8130-5435-3. [Google Scholar]
  17. Thompson, K.M.; Hood, A.; Cavallaro, D.; Lentz, D.L. Connecting Contemporary Ecology and Ethnobotany to Ancient Plant Use Practices of the Maya at Tikal. In Tikal: Paleoecology of an Ancient Maya City; Lentz, D.L., Dunning, N.P., Scarborough, V.L., Eds.; Cambridge University Press: Cambridge, UK, 2015; pp. 124–151. ISBN 978-1-107-02793-0. [Google Scholar]
  18. CONAGUA Servicio Meteorológico Nacional. Available online: https://smn.conagua.gob.mx/es/ (accessed on 25 January 2022).
  19. Holdridge, L.R.; Grenke, W.C.; Hatheway, W.H.; Liang, T.; Tosi, J.J.A.; WNRE INC CHESTERTOWN MD. Forest Environments in Tropical Life Zones. A Pilot Study; Defense Technical Information Center: Fort Belvoir, VA, USA, 1971; ISBN 978-0-08-016340-6. [Google Scholar]
  20. Pennington, T.D.; Sarukhán, J. Arboles Tropicales de México; Universidad Nacional Autónoma de México: Mexico City, Mexico, 2005. [Google Scholar]
  21. Miranda, F.; Hernández-X, E. Los Tipos de Vegetación de México y su Clasificación: Edición Conmemorativa 1963–2013; Ediciones Científicas Universitarias Serie texto Científico Universitario; Sociedad Botánica de México: Mexico City, Mexico, 2015; ISBN 978-607-16-1863-4. [Google Scholar]
  22. Martínez, E.; Sousa, M.; Ramos, C. Región de Calakmul, Campeche; Listados florísticos de México; Universidad Nacional Autónoma de México: Mexico, Mexico, 2001. [Google Scholar]
  23. Martínez, E.; Galindo-Leal, C. La Vegetación de Calakmul, Campeche, México: Clasificación, Descripción y Distribución. Bot. Sci. 2002, 71, 7–32. [Google Scholar] [CrossRef]
  24. Reese-Taylor, K.; Hernández, A.A.; Esquivel, F.C.A.F.; Monteleone, K.; Uriarte, A.; Carr, C.; Acuña, H.G.; Fernandez-Diaz, J.C.; Peuramaki-Brown, M.; Dunning, N. Boots on the Ground at Yaxnohcah: Ground-Truthing Lidar in a Complex Tropical Landscape. Adv. Archaeol. Pract. 2016, 4, 314–338. [Google Scholar] [CrossRef]
  25. Carr, C. Clasificación de Comunidades de Vegetación En Yaxnohcah Mediante La Utilización de Imágenes Satelitales de Landsat. In Proyecto Arqueológico Yaxnohcah, Informe de la 2016 Temporada de Investigaciones; Instituto Nacional de Antropología e Historia: Mexico City, Mexico, 2017. [Google Scholar]
  26. ESA. Sentinel Applications Platform; European Space Agency: Paris, France, 2022. [Google Scholar]
  27. Chave, J.; Réjou-Méchain, M.; Búrquez, A.; Chidumayo, E.; Colgan, M.S.; Delitti, W.B.C.; Duque, A.; Eid, T.; Fearnside, P.M.; Goodman, R.C.; et al. Improved Allometric Models to Estimate the Aboveground Biomass of Tropical Trees. Glob. Change Biol. 2015, 20, 3177–3190. [Google Scholar] [CrossRef]
  28. Zanne, A.E.; Lopez-Gonzalez, G.; Coomes, D.A.; Ilic, J.; Jansen, S.; Lewis, S.L.; Miller, R.B.; Swenson, N.G.; Wiemann, M.C.; Chave, J. Data from: Towards a Worldwide Wood Economics Spectrum. Ecol. Lett. 2009, 12, 351–366. [Google Scholar]
  29. Sullivan, M.J.P.; Talbot, J.; Lewis, S.L.; Phillips, O.L.; Qie, L.; Begne, S.K.; Chave, J.; Cuni-Sanchez, A.; Hubau, W.; Lopez-Gonzalez, G.; et al. Diversity and Carbon Storage across the Tropical Forest Biome. Sci. Rep. 2017, 7, 39102. [Google Scholar] [CrossRef] [Green Version]
  30. R Core Team. R: The R Project for Statistical Computing; The R Foundation: Vienna, Austria, 2022. [Google Scholar]
  31. Réjou-Méchain, M.; Tanguy, A.; Piponiot, C.; Chave, J.; Hérault, B. BIOMASS: An r Package for Estimating Above-ground Biomass and Its Uncertainty in Tropical Forests. Methods Ecol. Evol. 2017, 8, 1163–1167. [Google Scholar] [CrossRef]
  32. Bettinger, P.; Boston, K.; Siry, J.P.; Grebner, D.L. Valuing and Characterizing Forest Conditions. In Forest Management and Planning; Elsevier: London, UK, 2017; pp. 21–63. ISBN 978-0-12-809476-1. [Google Scholar]
  33. Slik, J.W.F.; Aiba, S.-I.; Brearley, F.Q.; Cannon, C.H.; Forshed, O.; Kitayama, K.; Nagamasu, H.; Nilus, R.; Payne, J.; Paoli, G.; et al. Environmental Correlates of Tree Biomass, Basal Area, Wood Specific Gravity and Stem Density Gradients in Borneo’s Tropical Forests. Glob. Ecol. Biogeogr. 2010, 19, 50–60. [Google Scholar] [CrossRef]
  34. Global Mapper Pro; Blue Marble Geographics: Hallowell, ME, USA, 2022.
  35. McGaughey, R.J. FUSION/LDV: Software for LIDAR Analysis and Visualization. Version 4.21; United States Department of Agriculture, Forest Service, Pacific Northwest Research Station: Portland, OR, USA, 2021.
  36. Fox, J.; Bouchet-Valat, M. Rcmdr-Package: R Commander; Chapman and Hall/CRC Press: Boca Raton, FL, USA, 2020. [Google Scholar]
  37. QGIS Geographic Information System; QGIS Association. 2022. Available online: https://qgis.org/en/site/ (accessed on 5 June 2022).
  38. Anaya Hernández, A.; Reese-Taylor, K. (Eds.) Proyecto Arqueológico Yaxnohcah. Informe de Las Temporadas de Investigación 2016; University of Calgary: Calgary, AB, Canada, 2017. [Google Scholar]
  39. Vázquez López, V.A.; Anaya Hernández, A.; Reese-Taylor, K. (Eds.) Proyecto Arqueológico Yaxnohcah. Informe de Las Temporadas de Investigación 2017–2018; University of Calgary: Calgary, AB, Canada, 2019. [Google Scholar]
  40. Read, L.; Lawrence, D. Recovery of Biomass Following Shifting Cultivation in Dry Tropical Forests of the Yucatan. Ecol. Appl. 2003, 13, 85–97. [Google Scholar] [CrossRef] [Green Version]
  41. Cairns, M.A.; Olmsted, I.; Granados, J.; Argaez, J. Composition and Aboveground Tree Biomass of a Dry Semi-Evergreen Forest on Mexico’s Yucatan Peninsula. For. Ecol. Manag. 2003, 186, 125–132. [Google Scholar] [CrossRef]
  42. NASA NASA JPL. Global Above Ground Biomass Mean Prediction. 2020. Available online: Earthdata.nasa.gov (accessed on 8 February 2022).
  43. Ortiz-Reyes, A.D.; Valdez-Lazalde, J.R.; Ángeles-Pérez, G.; los Santos-Posadas, H.M.D.; Schneider, L.; Aguirre-Salado, C.A.; Peduzzi, A. Transectos de datos LiDAR: Una estrategia de muestreo para estimar biomasa aérea en áreas forestales. Madera Y Bosques 2019, 25, e2531872. [Google Scholar] [CrossRef] [Green Version]
  44. Puc-Kauil, R.; Ángeles-Pérez, G.; Valdez-Lazalde, J.R.; Reyes-Hernández, V.J.; Dupuy-Rada, J.M.; Schneider, L.; Pérez-Rodríguez, P.; García-Cuevas, X. Species-Specific Biomass Equations for Small-Size Tree Species in Secondary Tropical Forests. Trop. Subtrop. Agroecosystems 2019, 22, 735–754. [Google Scholar]
  45. Lundell, C. Preliminary Sketch of the Phytogeography of the Yucatan Peninsula; Carnegie Institute of Washington: Washington, DC, USA, 1934. [Google Scholar]
  46. Lentz, D.L.; Hamilton, T.L.; Dunning, N.P.; Jones, J.G.; Reese-Taylor, K.; Anaya Hernández, A.; Walker, D.S.; Tepe, E.J.; Carr, C.; Brewer, J.L.; et al. Paleoecological Studies at the Ancient Maya Center of Yaxnohcah Using Analyses of Pollen, Environmental DNA, and Plant Macroremains. Front. Ecol. Evol. 2022, 10, 445. [Google Scholar] [CrossRef]
  47. Kabukcu, C. Wood Charcoal Analysis in Archaeology. In Environmental Archaeology: Current Theoretical and Methodological Approaches; Pişkin, E., Marciniak, A., Bartkowiak, M., Eds.; Interdisciplinary Contributions to Archaeology; Springer International Publishing: Cham, Switzerland, 2018; pp. 133–154. ISBN 978-3-319-75082-8. [Google Scholar]
  48. Hernández-Montejo, C. Del Palo de Tinte al Camarón; Gobierno del Estado de Campeche, Instituto de Cultura, Instituto de Antropología e Historia, Universidad Autónoma de Campeche: Campeche, Mexico, 2005. [Google Scholar]
  49. Bauer-Gottwein, P.; Gondwe, B.R.N.; Charvet, G.; Marín, L.E.; Rebolledo-Vieyra, M.; Merediz-Alonso, G. Review: The Yucatán Peninsula Karst Aquifer, Mexico. Hydrogeol. J. 2011, 19, 507–524. [Google Scholar] [CrossRef]
  50. Prümers, H.; Betancourt, C.J.; Iriarte, J.; Robinson, M.; Schaich, M. Lidar Reveals Pre-Hispanic Low-Density Urbanism in the Bolivian Amazon. Nature 2022, 606, 325–328. [Google Scholar] [CrossRef]
  51. Brewer, J.L.; Carr, C. Household Quarry-Reservoirs at the Ancient Maya Site of Yaxnohcah, Mexico. Lat. Am. Antiq. 2021, 33, 432–440. [Google Scholar] [CrossRef]
Remotesensing 14 03432 g001 550
Figure 1. Location of the study area in relation to the Calakmul Biosphere Reserve, SRTM elevation, UTM Zone 16N WGS 84.
Figure 1. Location of the study area in relation to the Calakmul Biosphere Reserve, SRTM elevation, UTM Zone 16N WGS 84.
Remotesensing 14 03432 g001
Remotesensing 14 03432 g002 550
Figure 2. Lidar point cloud showing the differences in vegetation heights, 1400 m long profile.
Figure 2. Lidar point cloud showing the differences in vegetation heights, 1400 m long profile.
Remotesensing 14 03432 g002
Remotesensing 14 03432 g003 550
Figure 3. Workflow of the vegetation classification and lidar-based AGB estimation model.
Figure 3. Workflow of the vegetation classification and lidar-based AGB estimation model.
Remotesensing 14 03432 g003
Remotesensing 14 03432 g004 550
Figure 4. (a) Vegetation classification of the Yaxnohcah study site over a Sentinel-2 satellite image. (b) Tree heights obtained from the difference between the Digital Surface Model and the Digital Elevation Model from lidar data.
Figure 4. (a) Vegetation classification of the Yaxnohcah study site over a Sentinel-2 satellite image. (b) Tree heights obtained from the difference between the Digital Surface Model and the Digital Elevation Model from lidar data.
Remotesensing 14 03432 g004
Remotesensing 14 03432 g005 550
Figure 5. Aboveground biomass for 73 transects (500 m2) per vegetation type.
Figure 5. Aboveground biomass for 73 transects (500 m2) per vegetation type.
Remotesensing 14 03432 g005
Remotesensing 14 03432 g006 550
Figure 6. Above ground biomass AGB Mg/500 m2. White area corresponding to wetland was excluded from AGB analysis.
Figure 6. Above ground biomass AGB Mg/500 m2. White area corresponding to wetland was excluded from AGB analysis.
Remotesensing 14 03432 g006
Remotesensing 14 03432 g007 550
Figure 7. Percentage of archaeological settlements located in each type of contemporary forest.
Figure 7. Percentage of archaeological settlements located in each type of contemporary forest.
Remotesensing 14 03432 g007
Remotesensing 14 03432 g008 550
Figure 8. Mound distribution of Yaxnohcah on topographic zones.
Figure 8. Mound distribution of Yaxnohcah on topographic zones.
Remotesensing 14 03432 g008
Table
Table 1. Aboveground biomass (AGB), aboveground carbon (AGC), and basal area (BA) at transects in Yaxnohcah, Campeche, Mexico.
Table 1. Aboveground biomass (AGB), aboveground carbon (AGC), and basal area (BA) at transects in Yaxnohcah, Campeche, Mexico.
Vegetation TypeNo. TransectsHaBasal Area
m2 ha−1		Sum of AGBAGB
Mg ha−1		AGC
Mg ha−1
Lowland211.0517.8584.0380.0337.70
Transition271.3525.19265.36196.5792.58
Upland241.2029.08289.05240.88113.45
Table
Table 2. Regression models for aboveground biomass (AGB) relating survey estimated AGB for three vegetation types and forest metrics calculated with airborne lidar data.
Table 2. Regression models for aboveground biomass (AGB) relating survey estimated AGB for three vegetation types and forest metrics calculated with airborne lidar data.
Vegetation TypeLinear ModelFAdj. R2RSE
LowAGB ~ 2.323 + 6.653 × MAD mode + −4.645 × AAD 15.180.590.77
TransitionAGB ~ 0.157 × exp(0.077 × “P95”) × “AAD” ^ 0.333 × 1.062 5.920.270.35
UplandAGB ~ 0.442 × exp(0.514 × “P80”) × exp(−0.298 × “P90”) × 1.03330.580.730.25
Table
Table 3. Topographic zone mound density.
Table 3. Topographic zone mound density.
Topo ZoneElevation (masl)f_Moundsf %Area (ha)Area %Density/ha
Bajo219.49–232.872928.121789.3133.240.37
Bajo Margins232.87–243.3484923.61866.0916.091.28
Mesoland243.34–252.59150241.771195.3722.211.43
Upland252.59–262.8185923.881114.1520.701.58
Highland262.81–281.54942.61418.287.771.83
Total 35961002410.511000.67
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Vázquez-Alonso, M.; Lentz, D.L.; Dunning, N.P.; Carr, C.; Anaya Hernández, A.; Reese-Taylor, K. Lidar-Based Aboveground Biomass Estimations for the Maya Archaeological Site of Yaxnohcah, Campeche, Mexico. Remote Sens. 2022, 14, 3432. https://doi.org/10.3390/rs14143432

AMA Style

Vázquez-Alonso M, Lentz DL, Dunning NP, Carr C, Anaya Hernández A, Reese-Taylor K. Lidar-Based Aboveground Biomass Estimations for the Maya Archaeological Site of Yaxnohcah, Campeche, Mexico. Remote Sensing. 2022; 14(14):3432. https://doi.org/10.3390/rs14143432

Chicago/Turabian Style

Vázquez-Alonso, Mariana, David L. Lentz, Nicholas P. Dunning, Christopher Carr, Armando Anaya Hernández, and Kathryn Reese-Taylor. 2022. "Lidar-Based Aboveground Biomass Estimations for the Maya Archaeological Site of Yaxnohcah, Campeche, Mexico" Remote Sensing 14, no. 14: 3432. https://doi.org/10.3390/rs14143432

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop