Did Drought Bring Down the Maya? — Setting the Record of the Weather Against the Record of the Cities

Stone Pillars That Stopped in the 9th Century

The low limestone land stretching across the southeastern Yucatán Peninsula of modern-day Mexico, northern Guatemala, and Belize is called the Maya lowlands. Villages and cities grew in this land starting before the Common Era, and from around the 3rd century onward, various kingdoms each ruled cities, allied, or fought one another. Classic Maya refers to Maya society from around 250 to 900 CE in the Maya lowlands, a period when kings ruled cities and erected stone pillars inscribed with dates. Pyramid temples and palaces rose in the northern Guatemalan city of Tikal, and kings carved accessions, wars, and rituals onto stone pillars along with their years. Yet in the 9th century, roughly 1,100 to 1,200 years ago, cities in the southern lowlands one by one stopped raising stone pillars and emptied out. Forests covered the temples, and modern people who rediscovered these ruins began to ask why this civilization ended.

In archaeology, collapse refers to a political order, such as a region's dynasty and cities, breaking down over a relatively short period. What happened to the people living within that order is not determined by this word alone.

In the case of the Maya, this story is often told in a single sentence. It is the sentence that the Maya people suddenly vanished into the forest. Similar sentences appear in academic journals. A 2007 paper on drought and the Maya collapse begins with the sentence, "Between 760 and 930 AD, millions of Maya disappeared from the face of the earth" (DOI: 10.1017/s0956536107000193). As records from lakes and caves have accumulated since the 1990s, drought has frequently been cited as the answer to this riddle (DOI: 10.1006/qres.1996.0042) (DOI: 10.1146/annurev-earth-060115-012512).

This article compares two sets of records. One is the weather records left at the bottom of lakes and in cave stalagmites, and the other is the city records left on stone pillars, house foundations, and the terrain beneath the forests. We examine in turn whether the dates of the two records overlap, and if they do overlap, what that means and where those meanings end.

What Lake Bottoms and Stalagmites Recorded About the Weather

There is a study that pulled a 6.3-meter-long mud cylinder from the bottom of Lake Punta Laguna in the northern Yucatán Peninsula (DOI: 10.1006/qres.1996.0042). Since no one was around 1200 years ago to measure how much it rained, the weather must be read from other things. Paleoclimate proxies are materials that, when there are no direct records of ancient weather, let us read whether that time was dry or wet by instead measuring traces left in nature that varied depending on the weather. A sediment core is a sample obtained by driving a long tube into the bottom of a lake or the sea and pulling up the mud piled in layers as a single column. Since deeper layers are older, reading the column from top to bottom goes back in time.

What was measured in the Punta Laguna core was oxygen in the shells of tiny crustaceans and snails that lived in the lake. Isotopes are atoms of the same element that differ slightly in weight. Just as "What Bones Tell" read ancient people's diets from the weight ratio of carbon and nitrogen, here the water conditions of the lake are read from the weight ratio of oxygen. When water evaporates, light oxygen escapes first, so in years with more evaporation than rain, the share of heavy oxygen in the lake water grows, and that share remains in the shells that grew in that water. Intervals in this core where the share of heavy oxygen is low indicate wet periods (DOI: 10.1006/qres.1996.0042).

Unusually dry events centered around 862, 986, and 1051 CE were read from that column, and this interval of frequent droughts was recorded in several other places in Mexico and Central America, overlapping in timing with the Classic Maya collapse (DOI: 10.1006/qres.1996.0042). Different lakes pointed to the same conclusion using different materials. When a drought shrinks a lake, gypsum dissolved in the water settles to the bottom, and a drought rhythm repeating roughly every 208 years was read from the oxygen and gypsum records of Yucatán lakes. The authors connect this cycle to changes in solar activity, and consider several of those drought peaks to correspond to the disruptions in Maya culture (DOI: 10.1126/science.1057759). In Lake Chichancanab on the same peninsula, ancient lake water trapped inside gypsum crystals deposited during droughts was directly analyzed, and the authors converted those values into models to estimate that annual precipitation dropped by 41 to 54 percent compared to today, and by up to 70 percent at its most severe (DOI: 10.1126/science.aas9871). Another analysis that interpreted four records together estimated that reduction at up to 40 percent, and considered it likely due to a decrease in summer tropical storms (DOI: 10.1126/science.1216629). Because the two numbers are estimates using different methods, they cannot be combined into a single value, but both point to "noticeably less rain than now."

The same period is read outside Maya lands as well. In the bottom sediments of the Cariaco Basin off the coast of Venezuela, the amount of titanium carried by rivers goes up and down following the region's rain, and in this record read on a seasonal scale, the collapse at the end of the Classic period sat in the middle of a prolonged dry spell, embedded within which were more severe droughts lasting several years, centered around 810, 860, and 910 CE (DOI: 10.1126/science.1080444).

Stalagmites are another clock. They are stone pillars formed as water droplets falling from a cave ceiling solidify layer by layer on the floor, so each layer preserves the oxygen of the rainwater from that time. Almost all Maya-region stalagmite studies use this oxygen weight ratio as an indicator of the rain that fell back then (DOI: 10.22201/cgeo.20072902e.2024.1.1751). The stalagmite from Yok Balum Cave in southern Belize recorded the past 2,000 years at intervals shorter than a year, complete with precise dating (DOI: 10.1126/science.1226299). A 2016 review that gathered and summarized these records concludes that the droughts from the 9th to 11th centuries were the most severe and frequent in the entire Maya history (DOI: 10.1146/annurev-earth-060115-012512).

There are also places where a single signal can be read in two ways. At Lake Salpetén in Guatemala, the share of heavy oxygen rose at the same time as the population decline between 800 and 900 CE, and the authors candidate both increasingly severe dryness and the effect of reduced water entering the lake as forests recovered as the causes (DOI: 10.1006/qres.2001.2305). Droughts were not even across regions either. In a 2024 review that laid records out on a map, the dryness signal of the same drought period was widely recorded in highlands and lowlands but did not appear everywhere in the Maya region (DOI: 10.22201/cgeo.20072902e.2024.1.1751).

Here is what to take away from this section. The fact that severe droughts occurred repeatedly in the Maya lowlands during the 9th to 11th centuries is a conclusion pointed to together by records from different places and different materials. Just how severe those droughts were differs in number depending on the method, and where and how much it rained varied by region.

What Overlaps and What Causes

The 2007 paper we saw at the beginning read four severe droughts around the years 760, 810, 860, and 910 in marine sediments from the Cariaco Basin, and interpreted that these four droughts lined up with the times when major cities were abandoned in four waves separated by about 50-year intervals (DOI: 10.1017/s0956536107000193). The more severe droughts singled out by the original study of the same basin are three around the years 810, 860, and 910 (DOI: 10.1126/science.1080444). The last stone pillar with a date carved into it at Tikal is from AD 869, and a period of unusually low rainfall continued into the mid-9th century (DOI: 10.1073/pnas.1408631111). The authors who studied Tikal's agriculture and water management conclude that a system already running at the limit of what the land could support failed to provide food, firewood, and drinking water in the face of a prolonged drought, which resulted in the collapse of social order and the abandonment of the city (DOI: 10.1073/pnas.1408631111). The authors of the Cariaco record write that a century-long decline in rainfall put a strain on resources, and a sudden drought amplified that strain, contributing to the social pressures that brought down the Maya (DOI: 10.1126/science.1080444).

All three conclusions start from the observation that the dates of the two records overlap. That droughts and collapses overlap in date is an observation; that droughts caused collapses is an interpretation built on top of that observation. For the interpretation to hold, the order at least must be right. Droughts must come first, and the dissolution must follow.

The results of comparing those sequences are mixed. That 2016 review points out archaeological evidence that the economic and political processes driving states to dissolution in the southern lowlands began before the major droughts. The same review clearly states the judgment that droughts nevertheless made the Classic collapse unusually severe and prevented the kind of recovery that seemed to have occurred in earlier periods (DOI: 10.1146/annurev-earth-060115-012512). At the city of Seibal in the Guatemalan lowlands, 154 carbon-dated ages were woven together with pottery studies to build a precise chronology, showing that warfare intensified around AD 735, multiple lowland cities collapsed around 810, and Seibal itself went into a major decline around 900 (DOI: 10.1073/pnas.1618022114). That method of dating ages using carbon remaining in charcoal and seeds is used in this sequence dispute.

The dates on the weather side also move. The researchers who first analyzed the Yok Balum stalagmite propose that unusually high rainfall from 440 to 660 helped population expansion and an increase in political centers, while drying from 660 to 1000 triggered the fragmentation of polities, an increase in warfare, and a staggered dissolution. They see the population collapsing amid a prolonged drought from 1020 to 1100 (DOI: 10.1126/science.1226299). A 2023 study that re-read the same stalagmite on a seasonal scale found that in AD 700 to 800, predicting when the rainy season would arrive became noticeably harder and severe droughts overlapped, and it was precisely in that period that the first dissolutions and population declines occurred. The authors suggest that abandonment was not driven by a single long-term drought alone, but that the increasing unpredictability of rain year after year made matters worse (DOI: 10.1038/s43247-023-00717-5). Depending on which record and which scale you look at, the answer to "Was the drought first?" changes.

There is a place where the two records sit side by side at a distance of 3km. The city of Uxbenká in southern Belize rose as the region's first political center after 100 BC, and the Yok Balum cave is located less than 3km away from its ceremonial center. This place is among the wettest in the Maya region, and with no traces of Classic-period wetland fields or large-scale agricultural investments, both food and drinking water depended entirely on rain. The researchers reconstructed how the city's population and construction investments grew and shrank using 167 carbon-dated ages and matched them against the rainfall record of the stalagmite right next door. The city actually grew rapidly after a severe drought in AD 200 and repeated growth and contraction until right after 750. When a severe drought began in 830, the political dissolution of southern Belize was already underway. The authors consider it likely that that 60-year drought contributed to the abandonment of Uxbenká and to preventing politics from reorganizing afterward (DOI: 10.1177/03091333221112359).

The order told by the two records in this single city is as follows. When the 60-year drought starting in 830 arrived, dissolution was already underway. However, a study that re-read the same stalagmite on a seasonal scale showed that there was also a severe drought earlier on from 700 to 800, and the first dissolution took place amidst that, so "already underway" is a statement based on the major drought of 830 (DOI: 10.1038/s43247-023-00717-5). The authors' judgment leans toward the idea that the drought of 830 helped along the abandonment of a faltering city and blocked politics from being reorganized.

The Paradox of Why the Drier North Lasted Longer

The northern Yucatán Peninsula receives much less rain today than the southern part, and it likely did so in the past as well (DOI: 10.1017/s0956536102132135). Yet while southern cities emptied out in the 9th century, many cities in the northern lowlands held on and even flourished under the hegemony of the northern Yucatán city of Chichen Itza (DOI: 10.1017/s0956536102132135). A 2002 paper addressing this paradox suggests that, in addition to the advantage of easy access to diverse resources, major shifts in the political and religious spheres were essential for Chichen Itza's adaptation. The authors read this case as a warning of the risk that the view that environment determines a society's fate falls into oversimplification (DOI: 10.1017/s0956536102132135).

The north eventually collapsed as well. A 2003 re-dated study links the process by which the Classic societal structure of the entire lowlands collapsed into a chain starting in the south over 200 to 250 years and ending with the 11th-century fall of Chichen Itza (DOI: 10.1017/s095653610314103x). That 2016 review notes that the late political collapse of the north around 1000 CE appears related to prolonged, severe droughts that followed (DOI: 10.1146/annurev-earth-060115-012512).

Different records offer different answers as to why the south collapsed earlier and more severely. Drought evidence up to that point came mostly from the drier and less populated north, which failed to explain why the wetter south was shaken earlier and harder (DOI: 10.1073/pnas.1419133112). A 2015 study measuring hydrogen and carbon in plant leaf surface waxes from sediment cores of two lakes in the north and south found that the south dried out more severely than the north at the end of the Classic period, and the researchers see this as matching the earlier and more persistent decline of the south. The researchers conclude that agricultural adaptations developed to match earlier droughts initially succeeded, but failed in the face of the more severe droughts at the end of the Classic period (DOI: 10.1073/pnas.1419133112). Unlike this study that directly measured two lakes, a 2024 review spreading existing records out on a map found that places that flourished during the collapse period were located in regions that experienced the worst droughts, while many places that were abandoned lay in water-rich regions. The authors review several interpretations together to explain this contradiction (DOI: 10.22201/cgeo.20072902e.2024.1.1751). Topography may have also been a fork in the road. An analysis suggests that the high ground of the peninsula's interior was more vulnerable to regime collapse and less suited to recovery than the surrounding lowlands (DOI: 10.1073/pnas.1114838109).

The fact that regions that lived through the same drought period met different fates means that the severity of the drought alone cannot explain who collapsed and who held on.

A City Revealed Beneath the Forest

In April 2009, an airplane scanned 200 km² over the Classic period city of Caracol in western Belize (DOI: 10.7183/2326-3768.2.3.208). Lidar is a surveying method that shoots many lasers from an airplane toward the ground and measures the time they take to return, filtering out only the light that reaches the ground through the leaves to create a topograph with the forest stripped away. Until then, the dense forest made it impossible to fully cover and survey ruins, and researchers could not be sure how well the samples they mapped and excavated represented the entire ancient settlement. In 2013, an additional 1,057 km² around it was surveyed, making it possible to view Caracol within its surrounding region and entire polity (DOI: 10.7183/2326-3768.2.3.208).

Caracol's map revealed dense residential areas, terraced fields, and reservoirs attached to the residential areas, where these terraced fields reduced soil erosion, retained water, and regulated rainwater flow (DOI: 10.7183/2326-3768.4.3.357). The research team that has studied this city for 30 years estimates that Caracol was home to more than 100,000 people during its peak between 600 and 700 CE (DOI: 10.1007/s10814-016-9101-z). This is the estimate produced by that research team for a single city.

In 2016, 2,144 km² of northern Guatemala was surveyed, revealing a landscape filled with interconnected urban settlements and large-scale infrastructure. The researchers write that these results prompt a reassessment of Maya population, agriculture, and political economy (DOI: 10.1126/science.aau0137). In a wetland in Belize, a wetland field complex previously known from remote and ground surveys turned out to be five times larger in reality, and an even larger complex was newly discovered (DOI: 10.1073/pnas.1910553116).

Headcounts do not come straight off these maps. A 2025 study gathered public and private lidar data and reprocessed even data captured for environmental surveys, estimating that 9.5 million to 16 million people lived across a vast area of the central Classic Maya lowlands, and presented a settlement distribution model spanning 95,000 km², a figure obtained by layering assumptions such as how many people lived in a single house onto the building traces picked up by lidar, and since lidar simply shows a snapshot of the present-day topography, whether those buildings were all used at the same time must be sorted out separately through excavation (DOI: 10.1016/j.jasrep.2025.105288) (DOI: 10.1371/journal.pone.0191619). A study analyzing 470 km² around Seibal sought to tackle precisely that limitation by overlaying architectural chronologies obtained from excavation onto the map, and calculating the rates at which building counts were overcounted or missed based on ground-verified results to calibrate the building counts and population estimates. The research team writes that these results too must be further tested and refined (DOI: 10.1371/journal.pone.0191619).

Ground truthing is the work of going to the site in person, walking, and digging to check whether the remotely detected traces are correct. Within 132 km² of lidar data from southern Belize, 315 groups of patio-enclosed house mounds were already known within a 35 km² area surveyed on foot, and visual scanning of the entire lidar map identified 563 more as "potential locations." The remaining 90 km² of that map is land that has not yet been surveyed on foot. Larger house mounds were more visible on the map, and the research team emphasizes that surveys walking on foot to verify remote findings remain necessary (DOI: 10.3390/rs12172838).

There is also concern that because lidar surveys were targeted around famous ruins, they inflate urban area and population. Researchers who analyzed lidar data captured in Campeche, Mexico, for environmental surveys unrelated to archaeology as a near-random sample argued that unrecorded cities and dense settlements appeared in that data as well, and that these were ubiquitous across the central lowlands (DOI: 10.15184/aqy.2024.148).

Lidar shifted the weight toward the view that the Classic lowlands were far denser than previously thought, and headcounts are numbers that must be heard together with their range and the assumptions used to build that range.

Was the Dense Land Strong or Weak Against Drought?

The landscape revealed by lidar makes one question read in two directions: Were dense fields and reservoirs devices to withstand drought, or structures that made them more dependent on rain?

Researchers in Tikal, northern Guatemala, present evidence that the Late Classic Tikal people fed a relatively large population with intensive agriculture—a mix of water delivery, terraced fields, tree tending, home gardens, and short fallows—along with a complex system for collecting and sharing water. The same research group sees a heavy price for that efficiency: it lowered environmental resilience and made them entirely dependent on reliably regular annual rains (DOI: 10.1073/pnas.1408631111). Wetland fields in Belize were in use during the exact period when population grew, the landscape changed, and droughts occurred, and the researchers view these fields as having been an adaptation to all those changes (DOI: 10.1073/pnas.1910553116). In other wetland research, most evidence points to wetland fields falling out of use at or shortly after the end of the Classic period—even including year-round waterlogged wetlands that should have been less sensitive to drought, with no traces of attempts to revive canals filled in after the Classic period (DOI: 10.1073/pnas.1114919109). This study also gathered evidence that dry periods overlapped with major transition periods. Still, considering simply that even fields with remaining water were abandoned, it is difficult to explain the reasons for abandonment solely by water scarcity.

The range of food sources was also evaluated. In a study classifying the drought tolerance of 497 native edible plants recorded as eaten by the lowland Maya, 83% were available during short droughts, but only 22% remained during medium droughts lasting up to a year, and 11% during multi-year extreme droughts. The authors premise that climate change alone cannot explain the multifaceted political turmoil of that period, and conclude that because the variety of foods besides maize was wide, food supply could have continued unless conditions were most extreme (DOI: 10.1073/pnas.2115657118). In a study reading diets from carbon and nitrogen weight ratios in the bones of 50 individuals from the ancient Maya community of Cahal Pech, the diet of the Late Classic upper class was narrow and skewed, and the authors consider that this may have made the socio-economic system more vulnerable, and that the system ultimately dissolved amid the landscape degradation and severe drought at the end of the Classic period (DOI: 10.1086/704530).

There is also the hypothesis that humans cut down forests and worsened droughts themselves. A 2012 study examined this hypothesis through climate model experiments. In simulations incorporating pre-Columbian land cover reconstructions, annual precipitation in southern Mexico and the Yucatán was 5–15% lower than when natural forest was included, and the authors propose that deforestation can explain up to 60% of the rainfall reduction during the Maya collapse period (DOI: 10.1029/2012gl052565). This figure differs from observations. It is a calculation within a model fed with reconstructions of how much the old forest was cut down, and changes when the input changes. Sometimes climate was the cause of forest reduction. Although long before the Classic period, the forest decline starting about 4,500 years ago in the record of Lake Petén Itzá in Guatemala began at the same time as a drying trend across the Caribbean region that lasted for about 1,500 years, and before human-cultivated maize pollen appeared. The researchers largely view the vegetation changes of about 4,500–3,000 years ago as the result of a dry climate (DOI: 10.1016/j.yqres.2008.10.002). Disentangling climate signals and human signals in lake sediments is a challenge in itself (DOI: 10.1017/s0956536102131063).

Conditions outside of drought also overlapped during the same period. A 2012 study covering the central lowlands considers that drying exacerbated the environmental pressure created by large-scale landscape change and resource demand, that overland trade crossing the peninsula shifted at the same time to a sea route going around the peninsula, and that as social conflict grew and elite control weakened, people chose to move to other parts of the peninsula rather than maintain a costly system (DOI: 10.1073/pnas.1210106109). In Caracol, artifacts left on floors just before abandonment point to warfare, the collapse of market systems, and heightened social tension (DOI: 10.1017/s0956536119000063).

A 2016 modeling study that simplified the feedback between water and society yielded results where even a slight reduction in rain caused an 80% drop in population. In the same simulations, population density and how sensitive crops were to drought were equally important, and while building reservoirs made the blows of drought rarer, when reservoirs dried up the blow was greater (DOI: 10.1002/2015wr018298). This result is a calculation showing the possibility that "even a small drought can invite a major collapse," and differs from observations that things actually happened that way in the Maya.

A dense landscape may have been both a device to withstand drought and a structure that made them more dependent on annual rain, and which factor was greater varied by place.

What Ended and Who Remained

In Seibal's timeline, the Classic collapse ended with the downfall of dynasties ruled by divine kings. Because the same city gave rise to such dynasties after surviving a collapse roughly 600 years earlier, the two collapses traced similar trajectories yet yielded different results (DOI: 10.1073/pnas.1618022114). In the central lowlands, forests largely returned after the sites were abandoned, and the population never came back (DOI: 10.1073/pnas.1210106109). Pollen records from Lake Petén Itzá also show forests reviving after the collapse around AD 900 (DOI: 10.1177/095968369600600302).

There is an interpretation that people in the central lowlands chose to move elsewhere on the peninsula rather than maintain an expensive system (DOI: 10.1073/pnas.1210106109), and cities persisted in the north. Chichen Itza endured until its decline in the 11th century (DOI: 10.1017/s095653610314103x), and afterward, an alliance centered on Mayapan rose and then fell in northern Yucatan (DOI: 10.1038/s41598-022-22756-2). When the Spanish army arrived in the 16th century, Maya city-states still existed in Yucatan and Guatemala. Today, millions of people in Mexico, Guatemala, and Belize still speak various branches of the Maya languages.

There is criticism within academia that the word "collapse" masks people's persistent resilience and the choices they made. A 2009 edited volume titled Questioning Collapse covered people's resilience and post-imperial life across multiple cases, including a chapter asking whether 21st-century worries were projected onto 8th-century Maya society (DOI: 10.1017/cbo9780511757815), and a 2017 primer argues that the collapse of ancient civilizations was far more complex than commonly acknowledged and that community resilience has been overlooked (DOI: 10.1017/9781316584941).

What ended in the southern lowlands in the 9th and 10th centuries were the dynasties ruled by divine kings and their cities; the Maya people and Maya languages continue to this day.

Answers Divided into Three Layers

The question "Did the Mayan civilization collapse because of a drought?" fits the evidence better when answered by dividing it into three layers rather than answering it in a single sentence.

• The fact that severe droughts occurred repeatedly in the Maya lowlands during the 9th to 11th centuries is a strong conclusion pointed to together by lake and ocean sediments and cave stalagmites.
• That the drought made the dissolution of the south more severe and prevented recovery is a judgment reached by several researchers. Answers still differ from record to record as to when and what came first, and how much rain fell in which region.
• The claim that the collapse can be explained by drought alone does not fit the current evidence. There were cities already shaking before the major drought of 830 (there are also records that the first dissolution occurred amid severe droughts earlier than that in 700–800), the north, which normally received less rain, held out longer, and wars, trade routes, and political changes overlapped at the same time.

The two records this article faces have different characteristics. Because stone pillars are records carved with the king's side of the story, we must first read who tried to leave what, as we saw in "Reading Historical Records — Grandmother's Diary and That Day's Newspaper". The dates of cities are dated using the method of "Dating with Carbon — Radiocarbon Dates, Read with an Hourglass and Tree Rings", and the dining tables of ancient people are read using the method of "What Bones Tell — One Person's Diary, Split Between Teeth and Bone". The more closely the dates of the two records are matched, the more accurately the gap between "overlaps" and "is because of" is measured.

Source Papers

• "Climate Variability on the Yucatan Peninsula (Mexico) during the Past 3500 Years, and Implications for Maya Cultural Evolution" (1996) — DOI: 10.1006/qres.1996.0042 — Punta Laguna core oxygen isotopes. Severe drying events and frequent drought intervals around 862, 986, and 1051 overlap with the Classic collapse.
• "Solar Forcing of Drought Frequency in the Maya Lowlands" (2001) — DOI: 10.1126/science.1057759 — Drought with a roughly 208-year cycle in Yucatan lake oxygen and gypsum records. Some peaks correspond to cultural disconnections.
• "Quantification of drought during the collapse of the classic Maya civilization" (2018) — DOI: 10.1126/science.aas9871 — Modeling Chichancanab gypsum hydration-water isotopes to estimate a 41% to 54% (up to 70%) reduction in annual precipitation.
• "Collapse of Classic Maya Civilization Related to Modest Reduction in Precipitation" (2012) — DOI: 10.1126/science.1216629 — Quantitative interpretation of four records, up to a 40% reduction, with a decrease in summer tropical storms as a possible cause.
• "Climate and the Collapse of Maya Civilization" (2003) — DOI: 10.1126/science.1080444 — Cariaco titanium. Multi-year droughts around 810, 860, and 910 amidst a long-term dry period. Author: contributed to social pressure.
• "Spatial and temporal distribution of palaeoclimatic records in the Maya Area" (2024) — DOI: 10.22201/cgeo.20072902e.2024.1.1751 — Stalagmites mostly use oxygen isotopes as a precipitation proxy. The paradox that drying signals are uneven across regions, with flourishing sites in strong drought areas and abandoned sites in water-abundant areas.
• "Development and Disintegration of Maya Political Systems in Response to Climate Change" (2012) — DOI: 10.1126/science.1226299 — Yok Balum stalagmite 2,000-year record. Suggests that wet conditions from 440 to 660 fueled expansion, while aridification from 660 to 1000 triggered fragmentation and collapse. Population collapse amidst long-term drought from 1020 to 1100.
• "Impacts of Climate Change on the Collapse of Lowland Maya Civilization" (2016) — DOI: 10.1146/annurev-earth-060115-012512 — Review. Droughts were most severe and frequent in the 9th to 11th centuries. The southern disintegration process began before the major droughts. Nonetheless, droughts contributed to the severity of the collapse and prevented recovery. Northern collapse around 1000 is linked to drought.
• "A 4000-Year Lacustrine Record of Environmental Change in the Southern Maya Lowlands, Petén, Guatemala" (2002) — DOI: 10.1006/qres.2001.2305 — Salpetén. An oxygen isotope rise between 800 and 900 may be due to aridification or reduced inflow from forest recovery.
• "DROUGHT AND THE MAYA COLLAPSE" (2007) — DOI: 10.1017/s0956536107000193 — "Millions of people vanished" between 760 and 930. Interpretation that Cariaco's four droughts coincide with four phases of city abandonment.
• "Forests, fields, and the edge of sustainability at the ancient Maya city of Tikal" (2014) — DOI: 10.1073/pnas.1408631111 — Tikal intensive agriculture and water management, the cost of lowered resilience and rainfall dependence. Last-dated stela 869, authors' conclusion that the system failed in the face of drought.
• "High-precision radiocarbon dating of political collapse and dynastic origins at the Maya site of Ceibal, Guatemala" (2017) — DOI: 10.1073/pnas.1618022114 — Bayesian analysis of 154 dates. Intensified warfare in 735, collapse of multiple centers in 810, Ceibal decline in 900. Classic collapse culminates in dynastic downfall.
• "Decline in seasonal predictability potentially destabilized Classic Maya societies" (2023) — DOI: 10.1038/s43247-023-00717-5 — First disintegration amid a decline in seasonal rainfall predictability and severe drought from 700 to 800. Suggests that reduced predictability, not just long-term drought, was an aggravating factor.
• "The development and disintegration of a Classic Maya center and its climate context" (2022) — DOI: 10.1177/03091333221112359 — Comparing 167 dates from Ucanal with a stalagmite 3 km away. Disintegration was already underway when the drought began in 830. A 60-year drought likely contributed to the abandonment.
• "CLIMATE CHANGE AND THE END OF THE CLASSIC PERIOD IN YUCATAN: Resolving a paradox" (2002) — DOI: 10.1017/s0956536102132135 — The drier north flourished under Chichen Itza during drought periods. Resource diversity and political-religious changes. Warning against environmental determinism.
• "THE NORTHERN MAYA COLLAPSE AND ITS AFTERMATH" (2003) — DOI: 10.1017/s095653610314103x — The lowland-wide collapse as a chain reaction proceeding from south to north over 200 to 250 years, ending with the 11th-century fall of Chichen Itza.
• "Drought, agricultural adaptation, and sociopolitical collapse in the Maya Lowlands" (2015) — DOI: 10.1073/pnas.1419133112 — Plant wax isotopes. The south was drier than the north at the end of the Classic period. Agricultural adaptation to earlier droughts failed under more severe droughts.
• "Kax and kol: Collapse and resilience in lowland Maya civilization" (2012) — DOI: 10.1073/pnas.1114838109 — The interior highlands of the peninsula are more vulnerable to collapse and less favorable to recovery than the lowlands.
• "The Use of LiDAR in Understanding the Ancient Maya Landscape" (2014) — DOI: 10.7183/2326-3768.2.3.208 — 200 km² of Caracol in 2009, 1,057 km² of lidar in 2013. Overcoming the limitation of not knowing sample representativeness due to forests.
• "Using Lidar and GIS to Investigate Water and Soil Management in the Agricultural Terracing at Caracol, Belize" (2016) — DOI: 10.7183/2326-3768.4.3.357 — Dense housing, terraced fields, and reservoirs. Erosion control, water retention, and runoff regulation by terraced fields.
• "Caracol, Belize, and Changing Perceptions of Ancient Maya Society" (2016) — DOI: 10.1007/s10814-016-9101-z — Summary of 30 years of research. Estimated population of over 100,000 living during the 600–700 heyday.
• "Ancient lowland Maya complexity as revealed by airborne laser scanning of northern Guatemala" (2018) — DOI: 10.1126/science.aau0137 — 2,144 km² of lidar in 2016. Demands reassessment of connected urban settlements, infrastructure, population, agriculture, and political economy.
• "Ancient Maya wetland fields revealed under tropical forest canopy from laser scanning and multiproxy evidence" (2019) — DOI: 10.1073/pnas.1910553116 — Wetland field complexes five times larger than previous estimates, with larger complexes newly discovered. Likely an adaptation used during population expansion and drought periods.
• "New regional-scale Classic Maya population density estimates and settlement distribution models through airborne lidar scanning" (2025) — DOI: 10.1016/j.jasrep.2025.105288 — Estimated 9.5 million to 16 million people across 95,000 km² of the central lowlands.
• "Archaeological application of airborne LiDAR to examine social changes in the Ceibal region of the Maya lowlands" (2018) — DOI: 10.1371/journal.pone.0191619 — Lidar provides only synchronic topography. Correction of building counts and population using excavation chronologies and ground-truthing error rates, requiring additional verification.
• "Detecting Classic Maya Settlements with Lidar-Derived Relief Visualizations" (2020) — DOI: 10.3390/rs12172838 — 563 potential candidates in addition to 315 pedestrian survey sites. Larger ones are more visible, ground verification needed.
• "Running out of empty space: environmental lidar and the crowded ancient landscape of Campeche, Mexico" (2024) — DOI: 10.15184/aqy.2024.148 — Examining concerns of sample bias with environmental survey lidar, arguing that dense settlement was universal in the central lowlands.
• "Wetland fields as mirrors of drought and the Maya abandonment" (2012) — DOI: 10.1073/pnas.1114919109 — Most wetland fields abandoned around the end of the Classic period, including permanently wet ones less sensitive to drought. No attempts at canal repair.
• "Large variation in availability of Maya food plant sources during ancient droughts" (2021) — DOI: 10.1073/pnas.2115657118 — 497 edible plant species: short droughts 83%, 1 year 22%, multi-year 11%. Food supply sustainable except under extreme conditions.
• "The Role of Diet in Resilience and Vulnerability to Climate Change among Early Agricultural Communities in the Maya Lowlands" (2019) — DOI: 10.1086/704530 — Isotopes of 50 individuals from Cahal Pech. Narrow diets among elites in the late Classic period may have increased vulnerability.
• "Pre‐Columbian deforestation as an amplifier of drought in Mesoamerica" (2012) — DOI: 10.1029/2012gl052565 — Climate model experiment. Suggests deforestation reduced annual rainfall by 5% to 15%, explaining up to 60% of the drying during the collapse period.
• "Climate drying and associated forest decline in the lowlands of northern Guatemala during the late Holocene" (2008) — DOI: 10.1016/j.yqres.2008.10.002 — Lake Petén Itzá. Forest decline roughly 4,500 to 3,000 years ago occurred before the appearance of maize pollen, largely as a result of arid climate.
• "PALEOLIMNOLOGY OF THE MAYA LOWLANDS" (2002) — DOI: 10.1017/s0956536102131063 — Difficulty of separating climate and human signals in lake sediments. 800 to 1000 was the late Holocene's driest period.
• "Classic Period collapse of the Central Maya Lowlands: Insights about human–environment relationships for sustainability" (2012) — DOI: 10.1073/pnas.1210106109 — Complex human-environment interactions, maritime shifts in trade routes, migration decisions. Environment recovers, population does not return.
• "FINAL MOMENTS: CONTEXTUALIZING ON-FLOOR ARCHAEOLOGICAL MATERIALS FROM CARACOL, BELIZE" (2020) — DOI: 10.1017/s0956536119000063 — On-floor artifacts before abandonment point to warfare, market collapse, and social tension.
• "Conceptualizing socio‐hydrological drought processes: The case of the Maya collapse" (2016) — DOI: 10.1002/2015wr018298 — Socio-hydrological model. Small precipitation reductions could cause an 80% population collapse, with density and crop sensitivity playing equal roles; the double-edged nature of reservoirs.
• "A Holocene vegetation history from lowland Guatemala" (1996) — DOI: 10.1177/095968369600600302 — Lake Petén Itzá pollen. Forest regeneration following the collapse around 900 CE.
• "Northeast Yucatan hurricane activity during the Maya Classic and Postclassic periods" (2022) — DOI: 10.1038/s41598-022-22756-2 — 2,200-year hurricane record. A period encompassing the declines of Chichen Itza and Coba and the rise and fall of the Mayapan League.
• "Questioning Collapse" (2009) — DOI: 10.1017/cbo9780511757815 — An edited volume questioning the collapse narrative. Resilience and post-imperial society.
• "Understanding Collapse" (2017) — DOI: 10.1017/9781316584941 — A primer arguing that collapse was more complex than commonly acknowledged and that community resilience has been overlooked.