Differential preservation can obscure activities at submerged archaeological sites. Wooden pole–and-thatch buildings associated with an ancient Maya salt industry are preserved underwater in mangrove peat at Ta’ab’ Nuk Na, one of the Paynes Creek Salt Works, Belize (McKillop and Sills Reference McKillop and Sills2022; Figure 1). Radiocarbon dating indicates three construction phases from the Early Classic to the Terminal Classic (McKillop and Sills Reference McKillop and Sills2022). Most buildings, including Buildings B, G, I, and J, were constructed during the Late Classic (AD 656–778). Building H dates to between the end of the Early Classic and the beginning of the Late Classic (AD 566–654). A sea-floor survey revealed abundant briquetage—pottery associated with boiling brine in pots to make salt—at Buildings G, H, I, and J, all of which are interpreted as salt kitchens. Building B is identified as a residence based on its domestic artifact assemblage (McKillop and Sills Reference McKillop and Sills2022).
Location of Ta’ab Nuk Na and other notable Paynes Creek Salt Works.
The sea floor consists of mangrove peat, an anaerobic but acidic sediment that preserves wood but not calcium in bone or other materials, depriving us of evidence of some activities (McKillop et al. Reference McKillop, Sills and Harrison2010). As a result, other types of evidence are necessary to reconstruct activity areas that are not preserved. In this study, chemical signatures obtained through inductively coupled plasma-atomic emission spectrometry (ICP-AES) provide insights into past activities associated with salt production, including brine boiling, brine enrichment, residences, storage, crafting, and fish preservation.
The associations of elements in soils and sediments derive from specific depositional processes. For example, phosphorus accumulates through the decomposition of organic materials such as bone, tissue, and plant remains, making it an indicator of food processing and disposal (Table 1; Middleton and Price Reference Middleton and Price1996). Ash derived from burned wood and plant materials enriches sediments with calcium, potassium, sodium, and magnesium (LeCount et al. Reference LeCount, Wells, Jamison and Mixter2016; Wells et al. Reference Wells, Novotny, Hawken and Glascock2007). Aluminum can reflect the clay content of the broken briquetage, while manganese may be influenced by repeated burning (Fernández et al. Reference Fernández, Terry, Inomata and Eberl2002). In the context of salt production, boiling brine and repeated heating and evaporation can leave residues enriched in these elements (Sills et al. Reference Sills, McKillop and Wells2016).
Elemental Associations Summarized from Previous Studies.
Table 1 Long description
Feature and activity areas are paired with elemental associations reported across prior studies. Enclosed spaces are linked to calcium and strontium in interiors, and to lower phosphorus where use is low. Food preparation and refuse are associated with phosphorus, magnesium, and manganese, sometimes with potassium and zinc. Fish processing, drying, and smoking are associated with sodium along with phosphorus, magnesium, and manganese. Sweeping and cleaning are characterized by low phosphorus. Brine boiling is linked to potassium, magnesium, aluminum, and manganese, while hearths are linked to potassium, magnesium, and calcium. Craft production and pigments are associated with copper, iron, manganese, lead, and zinc. These associations are summarized patterns and may vary by context, preservation, and sampling methods.
Sources: Anderson et alia Reference Anderson, Blair and Terry2012; De Lucia Reference De Lucia2013; Eberl et alia Reference Eberl, Álvarez and Terry2012; Fernández et alia Reference Fernández, Terry, Inomata and Eberl2002; Fulton et alia Reference Fulton, Wells and Storer2017; Hutson and Terry Reference Hutson and Terry2006; Knudson and Frink Reference Knudson, Frink, Hoffman and Price2004; Knudson et alia Reference Knudson and Frink2010; Lamoureux-St-Hilaire et alia Reference Lamoureux-St-Hilaire, Canuto, Wells, Cagnato and Barrientos2018; LeCount et alia 2015; Middleton and Price Reference Middleton and Price1996; Parnell et alia Reference Parnell, Terry and Nelson2002a, Reference Parnell, Terry and Sheets2002b; Sills et alia Reference Sills, McKillop and Wells2016; Terry et alia Reference Terry, Fernández, Parnell and Inomata2004; Wells et alia Reference Wells, Terry, Parnell, Hardin, Jackson and Houston2000, Reference Wells, Novotny, Hawken and Glascock2007.
This study interprets the spatial distribution of multiple elements as records of human and environmental deposition, instead of relying on one element to represent an activity (Middleton and Price Reference Middleton and Price1996; Terry et al. Reference Terry, Fernández, Parnell and Inomata2004; Wells et al. Reference Wells, Novotny, Hawken and Glascock2007). Chemical signatures are evaluated at Ta’ab Nuk Na inside and outside of wooden buildings along with the results of artifact analysis (McKillop and Sills Reference McKillop and Sills2022) to identify activities at the salt works. These results highlight how sediment chemistry can be used to interpret activity areas in submerged environments.
Methods
The marine sediment samples from Ta’ab Nuk Na were collected during the 2019 excavation season from the surface of the peat matrix using a stainless-steel spoon (Figure 2). No off-site background samples were collected. Instead, comparisons were made within the site and inside and outside of the wooden buildings. The samples were submitted to the LSU AgCenter in the Department of Agricultural Chemistry. Using ICP-AES, 21 elements were measured, including boron, calcium, copper, iron, magnesium, manganese, phosphorus, potassium, sodium, sulphur, zinc, aluminum, barium, chromium, cobalt, lead, molybdenum, nickel, selenium, arsenic, and cadmium. The concentrations of the elements were reported in parts per million (ppm). The results from the sediment chemistry were added to the map of the wooden architecture in ArcGIS Pro using a Kriged semivariogram linear map with the quantile output set at 0.5. The resulting files estimated the surface of unknown values from the known values and was useful for comparing the elements inside and outside of the wooden buildings and open areas.
Photograph of sediment collection in Building B at Ta’ab Nuk Na. Photo by Heather McKillop.
Results
Analysis of 95 sediment samples from Building B and transects through other buildings shows quantitative and spatial variability in the concentrations of elements (Table 2). Elements that were either too low for the instrument to detect (selenium, arsenic, and cadmium), showed little variation (chromium, cobalt, lead, molybdenum, and nickel), or were an indicator of the sea floor matrix (sulfur) were removed from further analysis.
Mean Values Representing Concentration for Each Group of Samples Corresponding to Discrete Buildings or Areas.
Table 2 Long description
Table 2 presents mean elemental concentrations for six activity areas across the site, including residential, salt kitchen, mixed-use, and adjacent outdoor contexts. Overall, the salt kitchen areas, particularly the Building J salt kitchen/mixed-use area, exhibit the highest concentrations of several elements, including aluminum, iron, potassium, and zinc, while sodium concentrations are elevated across all salt-production contexts. In contrast, areas north of Buildings E and J generally show lower elemental concentrations. Sulfur is notably highest in the Building B residential area. These spatial differences suggest that salt production and related activities influenced the chemical composition of sediments and soils across the site.
Kriged linear variograms show spatial variation in elemental concentrations inside and outside of the buildings (Figures 3–8; Table 2). Building B is different from other areas, with lower sodium concentrations; moderate values of boron, iron, magnesium, phosphorus, potassium, zinc, and barium; and high calcium, manganese, and aluminum. Buildings E, G, and H exhibit consistently elevated sodium and high boron, magnesium, phosphorus, potassium, and barium concentrations. Building J follows the same pattern, but copper, iron, and zinc concentrations are higher. Areas between buildings show moderate concentrations. The area north of Building J has lower concentrations of magnesium, manganese, potassium, zinc, aluminum, and barium. However, phosphorus concentrations are similar to those in Building B, and calcium concentrations are similar to those in the salt kitchens. In contrast, the area north of Building E shows lower concentrations compared to those in the other open areas.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in boron and sodium are shown using a Kriged semivariogram linear map in ppm. Wooden posts were recorded as either hardwood (buildings) or palmetto palm posts (linear lines as boundary features).
Figure 3 Long description
Two maps of Ta’ab Nuk Na show interpolated boron and sodium concentrations derived from Kriged semivariogram linear maps in parts per million (ppm). The left map displays boron concentrations classified into five ranges: 16-20, 21-22, 23-25, 26-30, and 31-40 ppm. The right map displays sodium concentrations classified into five ranges: 4,931-6,214, 6,215-6,993, 6,994-7,665, 7,666-8,200, and 8,201-8,826. Both maps identify Buildings B, E, G, H and J, hardwood posts, lines of palmetto palm posts, and sediment collection units. The maps highlight spatial variation in boron and sodium vconcentrations around the residence and salt kitchens. Both maps include a scale bar representing 14 m.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in aluminum and phosphorus are shown using a Kriged semivariogram linear map in ppm.
Figure 4 Long description
The two maps of Ta’ab Nuk Na display interpolated aluminum and phosphorus concentrations derived from Kriged semivariogram linear map in parts per million (ppm). The left map displays aluminum concentrations classified into five ranges: 5,064 to 6,948, 6,949 to 7,890, 7,891 to 8,680, 8,681 to 9,546, and 9,547 to 11,558. The right map displays phosphorus concentrations classified into five ranges: 16 to 27, 28 to 35, 36 to 45, 46 to 56, and 57 to 1110. S Both maps identify Buildings B, E, G, H, and J, hardwood posts, lines of palmetto palm posts, and sediment colleciton units. The maps highlight spatial variations in aluminum and phosphorus elemental concentrations around the residence and salt kitchens. Both maps include a scale bar representing 14 meters.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in potassium and calcium are shown using a Kriged semivariogram linear map in ppm.
Figure 5 Long description
Two maps of Ta’ab Nuk Na display interpolated potassium and calcium concentrations derived from Kriged semivariogram linear maps in parts per million (ppm). The left map displays potassium concentrations classified into five ranges: from 812- 885, 886-957, 958-1,031, 1,032-1,126, and 1,127-1,307. The right map displays calcium concentrations classified into five ranges: 1,081-1,540, 1,541-1,798, 1,799-2,042, 2,043-2,358, and 2,3592,911. Both maps identify Buildings B, E, G, H, and J, hardwood posts, lines of palmetto palm posts, and sediment collections units. The maps highlight spatial variations in potassium and calcium elemental concentrations around the residence adn salt kitchens. Both maps include a scale bar representing 14 m.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in magnesium and copper are shown using a Kriged semivariogram linear map in ppm.
Figure 6 Long description
Two maps of Ta′ab Nuk Na display interpolated magnesium and copper concentrations derived from Kriged semivariogram linear maps. The left map displays magnesium concentrations classified into five ranges: 1,4403-1,779, 1,780-,1921, ,1922-2,062, 2,063-2,236, and 2,237-2,606. The right map displays copper concentrations classified into five ranges: 7-13, 14-18, 19-23, 24-28, and 29-35. Both maps identify Buildings B, E, G, H, and J, hardwood posts, lines of palmetto palm posts, and sediment collection units. The maps highlight spatial variations in magnesium and copper elemental concentrations around the residence and salt kitchens. Both maps include a scale bar indicating 14 meters.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in iron and manganese are shown using a Kriged semivariogram linear map in ppm.
Figure 7 Long description
Two maps of Ta’ab Nuk Na display iron and manganese concentrations derived from Kriged semivariogram linear maps in parts per million (ppm). The left map displays iron concentrations classified into five ranges: 2,956-4,398, 4,399-5,120, 5,121-5,842, 5,843-6,671, and 6,672-9,772 ppm. The right map displays manganese concentrations classified into five ranges: 10-14, 15-17, 18-20, 21-24, and 25-33 ppm. Both maps identify Buildings B, E, G, H, and J, hardwood posts, lines of palmetto palm posts, and sediment collections units. The maps highlight spatial variations in iron and manganese elemental concentrations around the residence and salt kitchens. Both maps include a scale bar representing 14 m.
Map of Ta’ab Nuk Na showing wooden posts, buildings, and sediment collection units. Fluctuations in zinc and barium are shown using a Kriged semivariogram linear map in ppm.
Figure 8 Long description
Two maps of Ta’ab Nuk Na display interpolated zinc and barium concentrations from Kriged semibariogram linear maps in parts per million (ppm). The left map displays zinc concentrations classifies into five ranges: 6.3-8.2, 8.3-9.5, 9.6-11, 11.1-12.9, and 13-17.7. The right map displays barium concentrations classified into five ranges: 7-8, 9-10, 11, 12, and 13-15. B. Both maps identify Buildings B, E, G, H, and J, hardwood posts, lines of palmetto palm posts, and sediment collection units. The maps highlight spatial variations in zinc and barium elemental concentrations around the residence and salt kitchens. Both maps include a scale bar representing 14 m.
Discussion
Sediment chemistry proved useful in interpreting activities at Ta’ab Nuk Na, given the preservation bias in the mangrove peat (Table 2). Building B is chemically different from the other buildings, supporting the interpretation of a residence. Elevated calcium supports the interpretation of an enclosed space. Areas of iron, magnesium, potassium, and aluminum suggest a hearth in the northeast corner, likely used for food preparation (Table 1; Figures 3–8). This interpretation is supported by sea-floor survey data that show a diverse assemblage of artifacts in Building B, including serving vessels, chert tools, and a spindle whorl, which are absent in the salt kitchens (McKillop and Sills Reference McKillop and Sills2022:Figures 6 and 8).
Buildings E, G, H, and J resemble the chemical signatures associated with the salt kitchens at Chan b’i (Sills et al. Reference Sills, McKillop and Wells2016). The elevated magnesium, phosphorus, and potassium concentrations are consistent with brine boiling and fuel use. The elevated concentrations of manganese and aluminum reflect repeated burning and the breakdown of briquetage during production. Wood combustion produces ash, which contains calcium, magnesium, and potassium, although calcium appears less concentrated within the buildings than outside them. This is likely due to the sweeping and disposal of ash and material outside of the buildings. However, while Building J conforms to the salt production pattern in Buildings E, G, and H, its elevated concentrations of iron, zinc, and copper together with phosphorus, magnesium, and manganese, along with iron, zinc, and copper, may also be associated with craft production or fish processing (Table 1). Rather than representing a separate activity, this variation likely reflects differences in salt production intensity or a change in fuel use. Artifactual evidence supports the interpretation of this space as a salt kitchen due to the presence of briquetage inside the building (McKillop and Sills Reference McKillop and Sills2022:Figure 10).
The area associated with Building I differs slightly from the salt kitchens because of its elevated boron (likely due to a variation in the environmental conditions) and moderate phosphorus and calcium but lower concentrations of other elements. Sediment sampling began outside the building instead of inside, and this pattern may reflect activity areas adjacent to the building. The presence of clay funnel fragments associated with brine enrichment suggests that this area was used to concentrate brine prior to boiling (McKillop and Sills Reference McKillop and Sills2022:Figure 9f). Areas surrounding the salt kitchens show elevated concentrations of elements associated with salt kitchens consistent with the redistribution of production debris beyond the buildings.
The area north of Building J, where wooden stakes interpreted as supports for a drying rack were recovered, likely represents a related but separate activity to salt production, such as drying or preserving fish (McKillop and Aoyama Reference McKillop and Aoyama2018; McKillop and Sills Reference McKillop and Sills2022:Figure 8). There are lower concentrations of phosphorus, potassium, and zinc than in the salt kitchens. However, these concentrations are similar to those in Building B, suggesting another activity such as drying salted fish on a drying rack with off-site deposal of the viscera. In contrast, the area north of Building E exhibits generally lower concentrations and less variation across the elements, suggesting an area of low use.
Conclusions
Sediment chemistry using ICP-AES is an effective way to reconstruct activity areas at Ta’ab Nuk Na, despite the preservation bias in the mangrove peat. Chemical patterns indicate that Building B was a residence, with evidence of a hearth and food preparation. Buildings E, G, H, and J share a similar chemical signature, associated with salt production. The variation among these buildings suggests difference in intensity or an additional use for salting fish. Adjacent areas, including those near Building I, indicate brine enrichment. Other surrounding spaces reflect the redistribution of salt production debris and associated activities, such as drying or preserving fish. These results demonstrate the utility of using sediment chemistry for identifying invisible activities at submerged sites.
Acknowledgments
We thank Louisiana State University and the University of Texas at Tyler for institutional support. Fieldwork was conducted in 2019 under excavation (IA/H/2/1/19[04]) and export (IA/H/3/1/19[02]) permits granted to Heather McKillop from the Belize Institute of Archaeology. We thank the LSU AgCenter for sediment analysis and the field team: Cher Foster, Rachel Watson, John Young, and host family Tanya Russ, John Spang, and Lyra Spang.
Funding Statement
This research was funded by the National Science Foundation Grant 1826653 to the authors.
Data Availability Statement
The remaining sediment samples are stored in the archaeology laboratory at Louisiana State University, Baton Rouge. Digital documents are catalogued with the Underwater Maya Project materials at Louisiana State University and the University of Texas at Tyler.
Competing Interests
The authors declare none.
Open access
