Appreciation of the Maya area begins with an understanding of the landscape (West, 1964; White and Hood, 2004). The geological foundation of the Maya forest is a karst limestone platform that influences the spatial distribution of all resources. Local variations in limestone are expressed in drainage features and seasonal water distribution (Beach et al., 2009). The porous limestone absorbs rain, and local precipitation averages vary from 500 mm in the dry northwest Yucatan Peninsula to 4,000 mm in the far south. In the central area, where the majestic ancient cities of Tikal and El Pilar stand (Figure 1), rainfall ranges from 1,500 to 2,000 mm a year (West, 1964; White and Hood, 2004), with rainwater seeping down to form subterranean flows beneath the bedrock that determine access to surface water on the surface (Scarborough, 1993; Ford, 1996; Lucero, 2003; Ferrand et al., 2012; Šprajc et al., 2021). Water drains from the rocky hills, ridges, and escarpments to collect in scattered depressions, with an estimated 40% of the region comprising wetlands (Fedick and Ford, 1990; Schwartz, 1990; Dunning et al., 2002, 2020; Ford and Nigh, 2015). These variable environments provide the vital resources that influence the locations of ancient and modern settlements in the Maya lowlands (Beach, 1998; Dunning et al., 1998, 2009, 2012; Beach et al., 2002, 2006, 2018b; Ferrand et al., 2012; Liendo et al., 2014). Land cover over the limestone base depends on local climate, rainfall, and slope conditions, with upland forests replete with magnificent towering trees that thrive in the fertile shallow soils and are cultivated for a range of economic uses (Reina, 1984).

Understanding the traditional milpa system lies at the crux of debates about past and present Maya agriculture. Few Maya scholars recognize the effective role of fire in management (Hernández Xolocotzi et al., 1995; Eastmond and Faust, 2006; Nigh, 2008; Nigh and Diemont, 2013; see also Peters, 2000; Anderson, 2005; Abrams and Nowacki, 2020; Jurskis et al., 2020) or comprehend the systems’ integral role in shaping the landscape (see Conklin, 1957; Dove, 1983, 1993; Van Vliet et al., 2013), even as they negatively cast the milpa cycle as extensive “slash-and-burn” agriculture. The milpa system is based on a hand cultivated poly-cropping scheme that intimately manages microhabitats for water, drainage, erosion, and soil fertility while selecting and directing perennial succession processes (Chazdon, 2014; Ford and Nigh, 2015; see Robinson and McKillop, 2013; Lentz et al., 2021). Archeologists assumed that this system failed to produce sufficient food for growing populations in the ancient past and the intensification of this subsistence mode necessarily led to deforestation. The agroecological evidence for the Maya milpa-forest garden system tells an alternative story (Schwartz and Márquez, 2015).

Environmental and agroecological scholars point to the value of traditional agricultural practices and their relationships to the landscape (Barrera Vásquez et al., 1977; Gómez-Pompa and Kaus, 1990, 1999; Gómez-Pompa and Del Amo, 1994; Atran, 2003; Altieri and Toledo, 2011). These studies show how the Maya milpa system, with cyclic and strategic dry-season burns every ∼20 years (Hernández Xolocotzi et al., 1995; Faust, 1998, 2001; Terán et al., 1998; Ford and Nigh, 2014), is adaptive in the tropical forest ecosystem by reducing fuel loads and regenerating forests (see Altieri, 2002). Intensification is based on scheduled, skilled, and knowledgeable investments of manual labor with tools appropriate to the woodland landscape, not the plow and draft animals employed by European farmers.

In the context of Maya geography, contemporary traditional farming practices are distinct from so-called conventional agricultural practices (Sumberg and Giller, 2022). Modern conventional farming, developed with Western goals of uniformity to maximize profits, is based on monocropping arable land and cleared cattle pasture, with high external inputs to maintain uniform yields that rely on technology and labor efficiencies (Bray, 1994; Gold, 1999; Scarborough, 2003; Smith, 2005; Montgomery, 2007). Traditional farming developed to minimize risk on the basis of local human-environment relations, balancing family needs with local labor, skill, and scheduling (Dumond, 1961; Boserup, 1965; Reina, 1967; Netting, 1977, 1993; Stone and Downum, 1999; Schreinemachers and Berger, 2011).

Agroecologists, economic botanists, and evolutionary psychologists who work with traditional Maya agriculturalists and the plants of the Maya forest identify the milpa subsistence system as highly integrated with the environment (Nations and Nigh, 1980; Gliessman, 1982, 1992, 1998; Gómez-Pompa and Bainbridge, 1993; Terán and Rasmussen, 1994; Quintana-Ascencio et al., 1996; Atran, 1999, 2000; De Clerck and Negreros-Castillo, 2000; Anderson, 2003; Gómez-Pompa et al., 2003; Levy Tacher and Rivera, 2005; Campbell et al., 2006; Nigh, 2006, 2008; Campbell, 2007; Corzo Márquez and Schwartz, 2008; Ford, 2008; Ross, 2011; among many others). The milpa-forest garden cycle today has continuity with the past. Records from the conquest and colonial times show practices comparable with the present (Terán and Rasmussen, 1995; Graham, 2006; Altieri et al., 2012; Ford and Clarke, 2019; Sánchez-Suárez et al., 2021). This system, developed from ancient adaptations and historical practices (Roys, 1952; Farriss, 1984; Terán and Rasmussen, 1995; Alexander, 2006; Zetina Gutiérrez, 2007; Zetina Gutiérrez and Faust, 2011; Rushton et al., 2020; Slotten et al., 2020), presents an effective subsistence adaptation for ancient Maya land use.

The landscape created by the milpa cycle embraces infield home gardens and diverse, accessible outfields interspersed among secondary growth and mature, closed canopy forests. The field-to-forest cycle, described for the ethnohistoric and contemporary Maya (Roys, 1931; Villa Rojas, 1945; Hernández Xolocotzi et al., 1995; Zetina Gutiérrez, 2007; Cook, 2016; Evans et al., 2021; Ford et al., 2021), creates this patchwork and demonstrates how the mosaic landscape provides resources to fulfill daily requisites of food, condiments, fiber, oils, fuel, gum, furnishings, supplies, medicine, toys, construction materials for buildings, household utensils for cooking, spinning, baskets, and habitat for animals; in short, all the everyday household necessities (Fedick, 1996). If these resource strategies can be projected back in time (Morell-Hart et al., Forthcoming), the entire landscape, with soil characteristics, geological assets, and animal habitats, was part of environmental interactions undertaken to meet common human needs.

The human environmental imprint relates to permanent residential units within an asynchronous forest-to-field mosaic. This mosaic would hinge on complex infield home gardens covering less than one-half hectare (Netting, 1974; Fedick, 1992; Fisher, 2014), with approximately 70% of the infields in shaded orchards and 30% in sunny fields (Ford and Nigh, 2015). The legacy of this system is evident in the widely acknowledged biodiverse Maya forest today (Primack et al., 1998; Mittermeier et al., 2000; TNC, 2021). Use of the open field, like a forest gap (Hartshorn, 1989), initiates the cycle of regeneration from which the perennial forest products are nurtured, selected, planted, and grown (Ford, 2008; Lentz et al., 2014, 2018). Outfields are interspersed within the regenerating succession and mature forests, ensuring sufficient variety in land cover across the landscape, providing habitat for animals, construction materials for housing, as well as fruits and other useful perennial resources from the dominant plants of the Maya forest (Table 1).

Rather than creating a patchwork of shifting slash and burn maize fields and abandoned plots left to “fallow,” the diverse land use strategies of the contemporary Maya produce a varied horizontal and vertical vegetation landscape (Ford, 2020). The horizontal landscape mosaic includes fields, successional perennials, and mature canopy, while a variety of vertical growth habits exist in home gardens and in milpa fields. The inclusion of perennial seedlings and resprouts in the milpa hastens the regeneration of the plot when it transitions toward forest. Selection for utility as well as diversity keeps the milpa cycle viable in the long-term. The sun loving field crops give way to first low and then high shady forest over time. This land-use system disperses annuals and perennials, aiding in water conservation, building soil organics, and inhibiting erosion. Agricultural extension officers in Belize and Guatemala observe that erosion is regularly witnessed in plowed fields and rarely noted in milpa fields (Tzul, 2018, personal communication).

Vanguard trees and shrubs gain ascendency after 4 years of sharing the milpa field with forbs¹ and grasses. It is the natural regeneration cycle, not neglect, that encourages the prevailing annuals to give way to perennials. Far from uncontrolled, and certainly not fallow (defined as an unseeded plowed field), this succession-building phase requires constant attention to direct secondary growth and provide opportunities to improve livelihoods with useful plants and animals (Emery and Thornton, 2008, 2012; Bongers, 2017; Levis et al., 2017). The succession provides a complex space, with sun-seeking annuals retreating around the emerging shade trees; it is a stepping stone in the forest transition where the farmer can influence land cover to meet social and economic needs (Guariguata, 2017). The value of this second growth is only now being realized (Chazdon, 2014).

It is critical to appreciate that all the necessities of ancient Maya life derived from their tropical landscape. Understanding the flexibility of the milpa-forest garden cycle and its importance to living in the forest is fundamental. Agricultural practices developed over millennia, working with natural processes to meet immediate food needs with annual crops and selecting significant perennials for long term requirements (Palerm, 1967, 1976; Wilken, 1971, 1987; Nations and Nigh, 1980; Gómez-Pompa, 1987; Hernández Xolocotzi et al., 1995; Terán and Rasmussen, 1995; Gliessman, 2001; Schwartz and Márquez, 2015; cf. Conklin, 1957). The Maya milpa cycle is an intensive system based on skill and scheduling that maintains biodiversity, lowers temperature, conserves water, checks erosion, and builds soil fertility.

Water, a critical resource for plants and animals, in the Maya forest is distributed unevenly over the year (e.g., Kramer and Hackman, 2021). There are three distinct seasons that impact land use recognized by local farmers: the cool wet, beginning in November-December, called Yaax Pak’al or first planting; the dry season, beginning around March, known as Yaax K’in or first sun; and the warm wet period, beginning in May–June, called Noh Pak’al or principal planting (Bricker, 2017). It is important to understand that this annual precipitation scheme is related to distant influences from the north, south, and east. The Atlantic brings the cool wet period that coincides with winter in North America and is characterized by weather systems known as nortes. The initiation of the warm wet period is dependent on the northward move of the Intertropical Convergence Zone (ITCZ) and the intensity and frequency of the rains depend on the eastern development of storms and hurricanes. In between these distinct wet periods is the dry season that varies in length according to the persistence of nortes and hurricanes.

Deluges as well as droughts, from the farmer’s point of view, are not measured annually. In fact, annual scales miss key information and interactions that impact farmers and farming (Tuxill, 2004; Kramer and Hackman, 2021). While relevant at the regional scale, coarse-grained temporal and spatial precipitation data of annual and even seasonal generalities fail to characterize patterns critical to the Maya farmer (Kramer and Hackman, 2021). Farmers in the Central Lowlands expect the annual cycle to give two maize yields; the largest and most reliable from the May–June planting with its September–October ripening depending on the maize race chosen (Reina, 1984; Tuxill et al., 2010). As insurance, there can be a dry season planting opportunity, depending on observations of precipitation (Reina, 1967). Outfield plots are strategically placed to take advantage of variability in drainage and aspect. The remarkable repertoire of edible plants number nearly 500 (Fedick, 2020), and many of these edibles are drought tolerant, thriving with intermittent or scarce rain fall (Fedick and Santiago, 2022), including specific maize races (Tuxill et al., 2010; Fenzi et al., 2017).

Of course, with rainfall agriculture (Tuxill, 2004; Whitmore and Turner, 2005), rain is required, but too much rain is just as menacing as too little (Lundell, 1978). Bad years are not measured by annual rainfall peaks or troughs, but rather on the timing related to harvest. High rainfall at a time when maize is maturing is devastating yet anticipated from tropical storms and hurricanes (Kramer and Hackman, 2021). Critical to successful crop yields are: (1) the timing of dry periods, when fields are burned; (2) sufficient rain for crops to mature; but (3) not too much rain to flood fields and damage crops. The reality on the ground is that the timing of the rain is much more important than overall rainfall (Tuxill, 2004). Low annual rainfall delivered to meet crop needs is good, while high annual rainfall delivered at the wrong time will destroy crops and lay waste the harvest. Precipitation cannot be measured at a coarse scale.

The nature of the milpa cycle provides for diverse land cover types at any given time. Open fields are rarely larger than a hectare (Schwartz and Márquez, 2015; Kramer and Hackman, 2021) and are interspersed within the many stages of forest regeneration. This variety enriches soil fertility with organic matter and inhibits erosion. Traditional milpa farming by small holders works to minimize risk over the long term, and field systems are monitored closely, designed to conserve water and check soil loss, factors that farmers everywhere are well attuned to (Handelsman, 2021).

For the Maya case, there is controversy over the proxies of instability and erosion that are generally predicated on a view of land cover that pits maize fields against forest. Lands that are found around settlements at centers, like Tikal (Carr and Hazard, 1961), are managed without landesque investment. Managing soil on slopes is important. The use of terraces and drainages are examples where marginal lands are brought into the agricultural regime, yet these systems are rare. Indeed, water travels too fast over steep slopes and still water in basins need drainage for crops. It is obvious that, in local situations, there may examples of erosion (Dunning and Beach, 2000; Carozza et al., 2007; Wahl et al., 2007; Beach et al., 2008, 2018a,b, 2019). These specific pulses identified at base of slopes, wetlands, as well as the occasional uplands, need to be carefully assessed in terms of regional implications. Specifics where erosion is identified need to be carefully assessed in terms of regional implications and alternative explanations.

While situations may vary, we have some idea of worldwide erosion rates for conventional vs. traditional systems (Hooke, 2000; Montgomery, 2007). Rates for preindustrial farming are very low, and rarely result in serious erosion. Montgomery (2007), the gold standard for evaluating erosion, indicates conventional agriculture is much more prone to impact erosion rates than traditional systems. This concords with field observations (Figure 2).

Maya farmers, like subsistence farmers throughout the world, maintain diverse home gardens (Caballero, 1992; Campbell, 2007) along with several outfield plots, which leverage different resources and incorporate well-drained and poorly drained lands to hedge against uncertainties of weather and crop yields (Hernández Xolocotzi et al., 1995). The home infield is the base operation; it is the core that is coordinated with multiple outfields. The land use cycles of open field crops, secondary shrub and tree cover, and mature canopy forest gardens are long-term, enhanced resource-regeneration systems that run for a period of about 20 years (Ford and Nigh, 2015). With variability in rainfall and the seasonal cycle of precipitation, farmers today, and certainly in the past, favor established cropping in home infield plots for their most intensive labor investments, while dispersed outfield areas mitigate production challenges that vary with precipitation levels. Areas of emerging and mature perennials provide specific animal habitats and resources, from homes for bees to fruits and construction materials. This system builds food sovereignty for the farmer and security for the family, supporting the redistribution of resources that are required by the local community.

This infield-outfield pattern of land use (Netting, 1989) contributes to a complex, asynchronous landscape mosaic that shifts from forest to field and back again with each family, generation, century, and millennial cycle (Table 2). By working within the natural regenerative system, swidden in general, and the Maya agricultural cycle in particular, is integral to forest maintenance (see Conklin, 1954). As recognized by palynologists, the open and succession phases of the cycle favor wind pollination produced by abundant sun-loving annuals and perennials, which are well represented by forbs in the pollen rain (Islebe et al., 2018; Vela-Pelaez et al., 2018). Investments in long-term perennials of the mature phases of the cycle are evident in the archeobotanical data (Trabanino-Garcia, 2014; Morell-Hart et al., Forthcoming) and are supported by animal pollination, a process identifiable in specific local captures of pollen rain (Wahl et al., 2007; Dunning and Beach, 2010; Bhattacharya et al., 2011; Dunning et al., 2012; Islebe et al., 2015). Built on a basic 20-year field-to-forest system, the Maya milpa-forest garden cycle intervenes in the natural succession process (Chazdon, 2014) to direct the production of the landscape to meet human needs. In forestry terms, the opening for crops is equivalent to the forest gap; this artificially constructed gap provides diverse fields filled with forbs and grasses in a manner similar to natural gaps (Hartshorn, 1989; Fredericksen and Pariona, 2002). Today the evidence of past land use is the feral forest itself, reflecting the legacy of a highly domesticated landscape (Ross, 2011; Thompson et al., 2015; Ford, 2020; Armstrong et al., 2021; cf. Michon et al., 2007; Iriarte et al., 2020).

The Maya forest is recognized today as anthropogenic in origin (Gómez-Pompa and Kaus, 1992, 1999; Gómez-Pompa et al., 2003; Gómez-Pompa, 2004; Barrera-Bassols and Toledo, 2005; among others), an obvious consequence of ancient human interactions with the environment (Campbell et al., 2006; Ford, 2008; Ross, 2011; see Covich, 1978). In the Archaic period, archeologists have identified domesticated crops such as chile, maize, squash, and beans, which made up the harvest of mobile horticulturalists, as well as the presence of palm fruits and avocado native to the lowlands (McClung de Tapia, 1992; Pope et al., 2001; Piperno and Stothert, 2003; Smalley and Blake, 2003; Neff et al., 2006; Casas et al., 2007; Pohl et al., 2007). This early experimentation represented the incipient agroecology from which the milpa cycle arose.

Tracking vegetation changes through time in the Maya forest area is challenging and the pollen record overrepresents wind pollinated plants. Woody plant species are overwhelmingly pollinated by animals, with only about 2% relying on wind to spread their pollen over the landscape (Kellman and Tackaberry, 1997; Bush and Rivera, 1998; Bradley, 1999; Ollerton et al., 2011). While some animal pollinated trees are represented in the local pollen rain of the Maya area, they are restricted to plants with open flower structures (Bush and Rivera, 2001), making it difficult to infer the presence and abundance of mature canopy forest species (Bush, 1995; Bhattacharya et al., 2011). Given these conditions, the pollen data can be quite revealing, and an understanding of potential land use strategies embedded in the milpa cycle can help to provide alternative explanations for discussion.

Of the dominant trees of the Maya forest (Table 1), only Brosimum alicastrum, commonly known as ramon, is wind pollinated. Ramon does not stand alone. It may indeed be a good proxy for the entire forest, but it would represent, at minimum, the presence of the dominant plants of the Maya forest. This is corroborated with the growing archeobotanical data that demonstrate the use of the dominant trees in prehistory (examples include Turner and Miksicek, 1984; Trabanino, 2012; Thompson, 2013; Trabanino-Garcia, 2014; Morehart and Morell-Hart, 2015; Thompson et al., 2015; Dussol, 2017; Dussol et al., 2017, Dussol et al., 2021; Varela Scherrer and Liendo Stuardo, 2021). The standard paleoecological interpretation has taken ramon as the indicator of the presence of forest and its absence an indication of deforestation (Figure 5). Clearly, the focus on pollen to understand the vegetation landscape is problematic and must be considered carefully (Islebe et al., 2015; Torrescano and Islebe, 2015; Torrescano et al., 2019).

The pollen signatures reveal, at first, worldwide climate changes, and later the complex interactions as the Maya adapted to their environment. The rise in temperature and precipitation attending the Holocene Thermal Maximum is mirrored well with the steep rise of the megathermal Moraceae and Brosimum-type pollen. The precipitation regime from 8,000 to 4,000 BP is consistent and relatively stable in the context of an overall drying trend noted for the region (Haug et al., 2001; Wahl et al., 2014; see Ford and Nigh, 2015). This setting of predictability and continuity is an important factor with respect to human adaptation. There is variation in the pollen signals, particularly in the relationship between forest and forbs. Forests dominate, yet forbs fluctuate from well below 5% to more than 20%, at the same time grasses hover around 5% (Figure 3). From the standpoint of the low-density, mobile horticulturalists of the Archaic, this was a time when natural resources were relatively predictable, and their impact on the forest was negligible. This relationship would later change with a shift in climate that brought radical highs and lows in precipitation creating instability.

Major climatic extremes that impacted the Maya region beginning in the Early Preclassic, and continuing from 4,000 and 2,000 years ago, provided an environmental impetus for adaptations that is detected archeologically in terms of permanent residential structures and incipient civic architecture. Maya farmers had to develop flexible strategies to cope with the radical shifts and unpredictability that characterized this chaotic period (Figure 3), and the successful adaptations developed at this difficult time must have laid the foundations for the later developments of the Maya civilization.

Originally interpreted simply as a shift from forest to savannah some 4,000 years ago (Binford et al., 1987; Rice, 1996; Webster, 2002, see Figure 5), a detailed examination of the pollen taxa reveals a much more diverse picture, where savannah grasslands play no role at all (compare Figure 3 with 5). The interpretation of savanna is driven by the classifications of “disturbance” and masks the existing variability evident in the pollen records (Brenner et al., 1990, 2002; Curtis et al., 1998; Leyden, 2002). Unpacking the generic disturbance taxa into forbs and grasses shows clearly that herbaceous forbs dominate grasses (see Figure 3 drawn from data presented in Brenner et al., 1990, 2002; Curtis et al., 1998; Leyden, 2002), exactly what one would expect from the implementation of milpa cycle agriculture (Slotten et al., 2020; Table 2). The growth of settled populations, and their expansion across the forested region around 3,000 years ago, corresponds to this “disturbance” land-cover signature, marking the early development of complex Maya communities in Preclassic times (Table 3).

Proportions of forest proxies to forbs appear reciprocal. The increase in the contribution of forbs, reflecting gaps in the forest (compare Table 3 and Figure 3), corresponds to changes in the forest recorded by the drop in the Moraceae and Brosimum-type pollen. While the interpretations have become more nuanced, the general perception that deforestation persists based on original presentations (Binford et al., 1987; see Figure 5), and persists among Maya scholars, despite the noteworthy evidence that grasses maintain a relatively low level throughout the sequence. The overall plant diversity increased with the proliferation of annual and perennial forbs representing c. 50% of identifiable taxa (Curtis et al., 1998; Brenner et al., 2002; Leyden, 2002). The drop in Moraceae/ramon more likely reflects the tree diversification preferences of the Maya, favoring trees useful for their fruit, construction materials, and other productive uses (see Cook, 2016; Ford, 2020), which are largely invisible because they are pollinated by birds, bees, and bats. Their pollen would therefore be less likely to settle into lakes and be retrievable by coring operations.

The climatic challenges the Maya faced during the Preclassic were enormous enough to produce the creative cultural adjustments that later underwrote their civilization. The uncertainties would have provided new opportunities in interactions with the landscape including the use of fire (Schüpbach et al., 2015; Anderson and Wahl, 2016; see also Abrams and Nowacki, 2020). Deluges followed by droughts in short sequence expedited the development of the milpa-forest garden system that endures today. One can only speculate about the triumphs and defeats this process of trial and error produced over the erratic centuries of this climate chaos period. There is no doubt, however, about the ultimate achievement: Maya settlements steadily grew and civic centers continued to be founded. These settlements flourished and ensured the prosperity of the Classic Period (Table 3), which influenced the composition of the Maya forest into the present day.

On the threshold of the Classic Period, some 2,000 years ago, irregularities in precipitation decreased, and there was more climatic stability. Classic Period paleoenvironmental indicators are largely constant and predictable in terms of precipitation and pollen (Figure 4). Precipitation ranges are comparable to those of the Holocene Thermal Maximum, but the overall climate was dryer. The uncertainties of the preceding centuries left their mark, and the land use strategies began during that period of climate chaos formed a solid base for growth and development. The cultural elaborations characteristic of Maya civilization—monumental temples, hieroglyphic texts, astronomical observations, and growing and diversifying settlements—were a consequence of continuity. With a more predictable environmental setting and less uncertainty, investments could be expanded in the civic realms.

Just as the human imprint of the Classic Period Maya is reflected in the paleoecological record, so is the evidence of the “collapse.” The reciprocal relationship of the forest proxies, Moraceae and Brosimum-type pollen (ramon) to forb pollen reverses without delay (cf Figure 4): as one rises the other drops, and the pollen record soon indicates the unrestrained expansion of pioneering ramon, known today to be competitive in the well-drained uplands (Schulze and Whitacre, 1999). As with the rise of ramon 8,000 years ago as it expanded into open niches during the warm and wet Holocene Thermal Maximum, evidence after 1,000 BP suggests expansion of ramon into a new, but different, niche (Figure 3). This new niche of open gaps was likely the neglected monumental architecture, left unmaintained to deteriorate, which the pioneering and limestone-loving ramon quickly exploited after being dispersed by animals, such as bats (Medellin and Gaona, 1999; Valdés Bérriz, 2015). The increase of ramon at Maya sites recognized today (Puleston, 1968, 1983; Lambert and Arnason, 1982; see also Burn and Mayle, 2008) may be best explained by the absence of monument management and ramon’s pioneering adaptation to flourish in rocky soils (Reina, 1984), just the habitat presented by derelict temples and palaces. Forbs stay between 20 and 40% until European contact and drop drastically after 500 BP. The presence of forbs is an indicator of the persistent practice of the milpa cycle by populations that continued to farm in the woodlands, like the contemporary Lacandon, without the overburden of political elites.

A preoccupation with drought has plagued the story of Maya civilization. Rainfall agriculture does need rain, but good farmers have developed practices and strategies based on keen observation to work with natural processes (Wilken, 1987; see also Dumond, 1961; Reina, 1967; Terán and Rasmussen, 1994; Hernández Xolocotzi et al., 1995; Tuxill, 2004; Schwartz and Márquez, 2015). Management efforts at the microhabitat level and observations that couple experience, skill, and knowledge form the basis for flexible adjustments to climate conditions. The intimate, local scale of small holder operations minimize risk over the long term and allow planning across time with expectations for good and bad harvests (Kramer and Hackman, 2021).

Maya farmers attune land use to the need for water conservation and favor specific trees in milpa fields for their economic values as well as their water conservation attributes. They make every effort to stagger field locations so that there is always some vegetation cover to conserve water with shade from trees, shrubs, and forbs (Ford et al., 2021). Skilled forest gardeners avoid broad clearings that expose land to excessive evapotranspiration and promote natural shade cover, even in fields. Home orchards, managed succession, and mature trees are of further importance to this cycle while providing necessary natural resources for food, construction, and medicine among other necessities (Ford, 2020).

Given the three recognized seasons, farmers identify their crop vulnerability to rainfall variation during the August dry period, known as the canicula, and the late growing season when plants reach maturity. Annual precipitation rates impact harvests less than a prolonged canicula or heavy rain when crops are ready to harvest. These are particular moments in the farming cycles that can arise in August or along the harvesting continuum from October to April.

While data on changing annual and decadal rainfall provide essential information on general environmental conditions within which the ancient Maya were living, they do not provide the appropriate scale for understanding farming strategies. These are coordinated at the local level, and even at the scale of individual plots, where understanding drainage qualities, provision of shade, and land cover to manage the water cycle and conserve moisture, as well as improve soil fertility and check erosion, are necessary to ensure the harvest.

Drought and deluge are factors that will menace the farmer’s harvest. Too little rain at the onset of planting is just as bad as too much rain at the climax of the crop. Climate modeling today shows that, in the Maya area, temperatures are increasing and precipitation is decreasing, but such trends do not reveal the timing of the precipitation. Farmers, at the local level, hedge their bets by dispersing their fields in wet and dry areas to leverage the uncertainties of seasonal weather dynamics (Morell-Hart et al., Forthcoming). Plots that cycle into the perennial phases break up the landscape between fields and forests, which assists water management and erosion by providing varied land cover.

Much has been made of erosion as a consequence of forest clearing. The timing of the so-called erosion feature, “Maya clay,” is recognized in the Petén Lakes but not in other areas of the Maya lowlands. The relationship between deforestation and erosion as established through pollen reconstructions is ambiguous. The archeobotanical data, which demonstrate that a wide variety of woods were available to the Maya, need to be included in paleoenvironmental reconstructions (Sheets et al., 2012; Dunning et al., 2018; Morell-Hart et al., Forthcoming). Moreover, the timing of the proposed erosion events comes early in the development of Maya society, when population levels and density were very low. Rice and Rice (1990) estimate population at c. 43 persons per sq km at this time, which is about 20% of the estimates for the height of the Classic Period. Yet the interpreted erosion inexplicably diminishes with the growth and expansion of Maya settlements.

Montgomery (2007) provides a basis for evaluating erosion and a revealing review of global data on agriculture and erosion rates, which indicates that conventional agriculture — based on the Western model of arable monoculture — induces average soil movements between 1.5 and 3.94 mm per year. Contrast this with conservation agriculture, reflecting traditional systems of labor-based polyculture, which produce average movements of 0.08–0.124 mm per year. Conventional systems range from 19 to 32 times greater than the rates of the traditional systems today. The ancient Maya certainly impacted their local landscapes (e.g., Beach et al., 2006, 2008, 2018a; Dunning et al., 2012, Dunning et al., 2019; Wahl et al., 2019), but generalizing from the small populations of the Preclassic (Culbert and Rice, 1990), which relied entirely on stone tool technologies (Denevan, 1992) when facing the vigorous growth of tropical vegetation, it is difficult to accept “Maya clay” as major erosion feature. “Maya clay” in most prominent in Salpeten range from 6 to 7 m, requiring 1,700 years to accumulate based on erosional averages at the highest conventional agricultural rate of 3.94 mm/year. The highest traditional agriculture rates of 0.124 mm/year for the same length of time would accumulate only 0.21 m. The presumption that ancient Maya land use strategies from 3,000 years ago would have such an impact on the landscape needs to be directly tested.

In sum, viewing paleoecological data through the lens of traditional land use strategies can bring concordance to what appear as disparate lines of data. Pollen and precipitation data together provide the context for interpreting the successful Maya occupation of their forested landscape. In light of the new perspectives presented here, new challenges to accepted interpretations are clear. What data will indicate the transition from mobile horticulturalists to settled agriculturalists? How did the forest garden management system transform social hierarchies and fuel the development of Maya civilization? And, curiously, what were the economic, political, and social upheavals that caused the famous “collapse?” Maya farmers, who understood the art of not being governed, were there at the time of the conquest, so population abandonment provides no adequate explanation.