A Summary of 25 Years of Research on Water Supplies of the Ancestral Pueblo People

Abstract

Six ancestral Pueblo community water supply sources were investigated by a team of engineers, scientists, archeologists, and other specialists affiliated with Wright Water Engineers, Inc. (WWE), and the Wright Paleohydrological Institute (WPI) from 1996 to 2021. The team members applied their various technical backgrounds and research methods to gain more insight into the water available to the ancestral Pueblo people living in the Four Corners area of the United States between 750 and 1280 CE, and how these indigenous people managed the water. Using lab analyses, field research, surveys, and analyses of sediment layers, the WWE/WPI team determined that four mounded areas discovered at Mesa Verde National Park had been ancestral Pueblo reservoirs. Through climate research, lab analyses, and investigations at these and two other sites, the team learned that water in this region was limited, and the community had to work diligently to harvest this water and maintain access to it. In the case of the four reservoirs studied, for example, the runoff used as water supply carried a high volume of sediment that required the water storage basins to be frequently dredged to maintain adequate capacity. These and other examples indicate that the ancestral Pueblo people were resourceful, hardworking, and organized water harvesters.

1. Introduction

As early as 7500 BCE, nomadic Paleo-Indians chose the area now known as the Four Corners region of the United States to be their home (Figure 1). The Four Corners area, where Utah, Colorado, Arizona, and New Mexico meet, accommodated six eras of ancestral Pueblo people: Basketmaker I, Basketmaker II, Basketmaker III, Pueblo I, Pueblo II, and Pueblo III. For about 750 years, the ancestral Pueblo people lived at Mesa Verde, first as nomadic hunters, later as farmers living in pit houses (Figure 2), and eventually as builders of remarkable cliff dwellings in the sheltered alcoves of the canyon walls (Figure 3) [1].

Water scarcity was a defining factor in the later ancestral Pueblo inhabitation of the Four Corners area. In the best of times, the climate was arid and water was a precious resource to be carefully managed. The worst of times brought a major drought that lasted from 1135 to 1180 CE. This was about the time that the ancestral Pueblo people moved from pleasant valley bottoms and mesa tops to cliff dwellings that could be better defended. Beginning around 1140 CE, the residents of the Four Corners area experienced famine, water uncertainty, and raiding parties. People began leaving Mesa Verde around 1250 CE to resettle along the Rio Grande, in southern Arizona and in Old Mexico. From 1275 to 1300 CE, there was another major drought; it was during this time that most of the remaining ancestral Pueblo people either abandoned the region or were killed by hungry raiders [2].

Mesa Verde National Park (MVNP), established in 1906 by Congress as a national park and ranked by the National Geographic Society in 1999 as number six of all world wonders, encompasses 210 square kilometers of the Four Corners area. The Pueblo I, Pueblo II, and Pueblo III people who lived in what is now MVNP and the surrounding area proved themselves to be successful in implementing water works and civil engineering projects that made life possible in the dry climate of southwestern Colorado. Wright Water Engineers, Inc. (WWE), and the Wright Paleohydrological Institute (WPI) collaborated on the study of water supply development throughout these three ancestral Pueblo eras. The multi-disciplinary WWE/WPI team explored reservoirs at MVNP that were later distinguished collectively as one of five American Society of Civil Engineers National Historic Civil Engineering Landmarks in Colorado, plus two other ancestral Pueblo water supply artifacts in the region—a cistern and a spring. The four reservoirs studied are Morefield, Far View (aka Mummy Lake), Sagebrush, and Box Elder, all described in detail in The Water Mysteries of Mesa Verde [3]. The WWE/WPI team also performed water supply research at Mug House cistern (MVNP) and Goodman Spring (Hovenweep National Monument). The data and analyses developed through this research were catalogued and made available to the public at the Norlin Library at the University of Colorado in Boulder, Colorado, in 2024.

The purpose of this manuscript is to provide a “one-stop” summary of WWE/WPI’s water research in the southwestern United States and the resultant findings. No other paper by the WWE/WPI team provides the overview and synthesis towards which this manuscript strives. This manuscript is intended to make other researchers aware of the studies conducted at these six sites to date, what was learned, and where further data can be gleaned. A larger purpose of this research is to look at the prehistorical water handling practices of a culture—the ancestral Pueblo people—to understand the importance of water as a resource and its role in shaping history. In the face of climate change, shifting resources, and evolving technology, learning about the extensive effort made by the ancestral Pueblo people to secure limited amounts of water should be both inspiring and prescient. While modern problems and solutions differ from those of the ancestral Pueblo people, their resourcefulness and determination can serve as a positive example for coming generations.

2. Materials and Methods

Webster’s Third International Dictionary defines paleohydrology as “the study of ancient use and handling of water” [4]. Research by the WWE/WPI team in the Four Corners area has been aimed at that definition. During a 25-year period, this team of water resource engineers, hydrologists, scientists, geologists, archeologists, palynologists, and anthropologists sought to learn about the water-harvesting practices of the ancestral Pueblo people through the six studies described herein. This work was intended to harness the expertise of veteran scientists from various fields to address a specific query—how did the ancestral Pueblo people of the southwestern United States obtain water and what can we learn about their practices and results?

The team obtained permits from the National Park Service (NPS) and followed NPS’s strict field protocols for preservation and protection of these sites. In some cases, the team collaborated to combine resources and data with archeologists who were already conducting excavations or studies.

The analyses typically applied at each site included procedures that, when integrated, would validate the purpose and function of each site and provide evidence of how the water structure was developed, how and when it was used, and the likely characteristics of operation and maintenance. These analyses included the following:

• Field topographic surveys using traditional civil engineering technology with theodolites, level instruments, measuring tapes, and planetables. The team defined and mapped the topography so that shapes of features, heights of mounds, and lengths of canal routes could be calculated (Figure 4).
• Using soil augers to recover soil profile data from the archeological sites so sediment layering could be defined with specificity, along with charcoal evidence. Soil profile data also helped define the physical extents of reservoir and canal building where natural, undisturbed soil layers were confirmed (Figure 5).
• Surface infiltration tests for site-specific data on present-day infiltration rates of bare and vegetated soils and typical permeability. These data informed estimates of rainfall–runoff relationships (Figure 6).
• Geologic descriptions for an understanding of basin characteristics and the source of eroded materials. The team defined bedrock outcrops and faults.
• Geomorphological analyses to define the character of canyon bottoms.
• Ceramic analyses performed by archeologists to estimate the period of human activity for reservoir dating.
• Paleo climate evaluations using dendrochronology (tree rings) for estimates of precipitation and temperature going back to about 500 CE. Overall, the long-term precipitation in the Four Corners area has been roughly equal to modern times, i.e., about 46 cm per year [5].
• Rainfall–runoff determinations to verify that a surface water supply existed.
• Pollen analyses of various sediments to define prehistoric vegetation, the type and location of agricultural practices, and the presence or absence of wetland plants. Forest successional periods and the presence or absence of sagebrush and medicinal plants were also evaluated.
• Laboratory analyses of reservoir sediments using hydrometer and sieve analyses of soils and sediments to define sand, silt, and clay content to estimate sediment transport and depositional character. The team looked for redoximorphic features that indicate long-term soil saturation.
• Carbon dating, where feasible, to determine the age of prehistoric structures and artifacts, such as a deer antler, to supplement pottery analyses.
• Groundwater evaluations to estimate water sources available and whether reservoir inflow might have originated via aquifers.
• Archeological studies of the reservoir structures, canals, watersheds, and adjacent villages, allowing the engineering data and studies to be properly placed in their anthropological context.
• Aerial, standard black and white, color and infrared photographs to define the ground conditions, wetlands, and water tables, and assist with the evaluation of geomorphological processes.

All or a combination of the above activities were conducted for the six sites summarized in this paper. The following sections highlight new information gleaned from the study of each site.

3. Discussion and Results

3.1. Morefield Reservoir

Morefield Reservoir, on the floor of Morefield Canyon, was the first ancestral Pueblo artifact studied by the WWE/WPI team after a MVNP superintendent asked for the team’s opinion on what this relic mound might have been. Several theories were in play, including that the mound could have been a ceremonial dance platform or perhaps a former reservoir (Figure 7). Through the processes of excavation and examination of mound layering (Figure 8), as well as soil testing, the team determined that the mysterious mound had originally been a 1.22-meter-deep, 15-m-diameter storage system for runoff and groundwater circa 750 CE. The evidence of this former use was fulsome: (1) the mound was comprised of sediment layer patterns consistent with the frequent dredging of waterlogged deposits that had been washed into the basin from the silty terrain upstream; (2) core samples extracted using soil augers indicated the site had been a tadpole-shaped indentation into the earth up to 6.5 m deeper than the highest part of the mound (Figure 9); (3) the sediment layers contained a concentration of shards from broken water vessels dating between 750 CE (at the deepest) and 1100 CE (at the shallowest) (Figure 10); (4) the soil analyzed from the excavation had redoximorphic features, indicating it had been submerged in water for long periods; (5) adequate runoff would have existed to occasionally fill this reservoir; and (6) the sediment contained high concentrations of maize pollen that were likely deposited by runoff from uphill agricultural fields.

This hand-dug basin captured runoff via an inlet canal (the tail of the tadpole shown on Figure 9) established by the ancestral Pueblo people around 750 CE. The team found evidence that the canal had been lined and relined with flat rocks to reduce erosion. Its length was extended over time as the erosive, silty soil that had washed in required continual dredging, creating rising side berms where the dredged material was thrown. WWE/WPI analyses determined that the silty sediment washed into the reservoir from the basin above would have had to have been dredged almost continually. Despite the cleaning operations, the valley-bottom reservoir and canal rose in elevation about 1.8 cm per year. After the reservoir was abandoned, the silt in the reservoir accumulated within the raised berms until the feature became the mound we see today.

The excavated mound layers contained more detailed information on how the reservoir developed. Strata of hardpacked sand indicated long-ago wind direction (Figure 11), while mangled layers indicated a berm failure in a section that was dug at too deep a slope. Maize pollen accumulation in the sediment layers indicated that the runoff filling the reservoir flowed through upstream corn crops. Carbon deposits in the sediment layers demonstrated that 14 forest fires had occurred during the reservoir’s 350-year period of use.

The Morefield Reservoir would have filled sporadically due to large rainfalls and increased runoff from periodic forest fires. Additional WWE/WPI team analyses of capacity and weather patterns suggested the Morefield Reservoir stored up to 450,000 L of water about five times per year, or about 2.25 million liters of water per year. While the WWE/WPI team did not specialize in anthropology, they were curious about the rough amount of water available per person in the Ancestral Pueblo era. According to Wilshusen [6], the average population of Pueblo I communities was about two hundred people. Dividing the estimated 2.25 million liters of water stored per year at Morefield Reservoir by 200 people, divided by 365 days per year, indicates that the residents of Morefield Reservoir might have had available approximately 30 L of water per person per day for drinking, cleaning, livestock watering, and other uses. This likely would have been a best- case scenario because Morefield Reservoir was the largest Four Corners water storage feature studied by the team. For comparison, Americans in 2015 used an average of 310 L of water per day [7].

The main thing that analysis of the Morefield mound layers taught the WWE/WPI team was that the ancestral Pueblo people were willing to engage in near-constant maintenance of this reservoir. Clearly, the water must have been regarded as a precious resource to merit this investment of extensive time and effort. Based on their various specialties, the WWE/WPI team members also concluded the ancestral Pueblo people must have been well organized and communal to successfully execute these reservoir-building and maintenance operations.

3.2. Far View Reservoir

The next reservoir studied by the WWE/WPI team was Mummy Lake, later renamed Far View Reservoir. In contrast to Morefield Reservoir’s valley-bottom location, Far View Reservoir is located on the top of Chapin Mesa (Figure 12). Far View Reservoir is one of about three dozen MVNP archeological sites with public access and is therefore quite popular. The former use of the site was particularly enigmatic because it is comprised of stone walls that make the structure appear significant, and perhaps ceremonial. In response to archeologists’ questions on whether there had been a water supply at Far View Reservoir, the WWE/WPI team conducted hydrological and scientific studies to analyze water supply feasibility and possible reservoir operations. The reservoir structure shown in Figure 12 is about 28 m in diameter and would have contained a maximum depth of 1.4 m, representing maximum storage of about 300,000 L of water. Sand and silt in the reservoir excavation were clearly deposited by water.

Palynological analyses of soil samples taken upstream of the reservoir found high concentrations of maize pollen. This suggested to the team that agricultural fields would have existed above the reservoir, allowing for greater runoff than would have flowed from undeveloped land. The team analyzed the soil inside of the stone walls outside of the water storage area and found higher levels of maize pollen there. The team developed a theory that if the reservoir stored water, like Morefield Reservoir, there would be a great deal of sediment running into it. The logical place for the dredged sediment to have been disposed would have been within the perimeter stone walls. The presence of high levels of maize pollen in this area suggested to the team that runoff from the corn fields above would have flushed maize pollen along with the sediment, which would have been thrown into the walled ring during dredging. This theory was borne out by archeologist David Breternitz, who analyzed sediment layering and pottery sherds to determine that the reservoir began as a simple hole in the ground with the walls added later, likely to hold the dredged material. A study of the material within the walls provided further evidence that Far View was a water storage feature, in the form of potsherds and the presence of pollen that indicated water-loving plants had grown there.

Surveying by the WWE/WPI team found that Far View Reservoir lies about 95 m in elevation below the highest point of the ridge. By analyzing weather patterns and typical runoff rates for agricultural and hard-packed ground conditions, the WWE/WPI team concluded that the site was indeed a reservoir that could have received runoff from interceptor ditches along the ridgeline, foot-packed public areas, and farmland, above. The reservoir has been known to collect water during modern times in much the same way it did during ancestral Pueblo times (Figure 13). The reservoir would not have received any groundwater and would have held surface water infrequently.

3.3. Sagebrush Reservoir

The WWE/WPI team also performed research at Sagebrush Reservoir (Figure 14), which resides 1.6 km west of Morefield Reservoir on an unnamed mesa. Sagebrush Reservoir is another mesa-top water storage facility, utilized by the ancestral Pueblo people from 950 to 1100 CE. A plan of the site can be seen in Figure 15. A WPI colleague, Dr. Jack Smith, had excavated Sagebrush Reservoir in the 1970s and found that it began as a dug-out water storage facility, without walls, and would have held approximately 340,000 L of water. Due to about 1.5 cubic yards of sediment accumulation per year, the capacity of the reservoir decreased over time and walls were constructed to retain the dredged material. The reservoir’s capacity was ultimately reduced to roughly 150,000 L by the time it ceased to be used as a water storage facility in about 1100 CE. Based on early water-borne deposits in the reservoir, followed by a period of eolian deposits (windblown sand, loess, and long-range-transported dust), the WWE/WPI team determined that the reservoir had been used by the ancestral Pueblo people for other purposes during its final stage of occupation. This conclusion was further confirmed by evidence of human activity, such as a metate fragment, found in the upper layer of deposits.

No water could have flowed to Sagebrush Reservoir without human intervention. Evidence was found on the ridge of several interceptor ditches beginning at higher elevations that routed runoff to the reservoir. As was the case with Far View Reservoir, Sagebrush Reservoir relied on hard-packed sand and clay terrain above it to produce runoff. The WWE/WPI team determined that about ½-acre of this water-tight ground surface could produce runoff during periods of about 1.3 cm or more of rain per hour. Based on precipitation data, the reservoir would have stored water about five or six times per year. A small spring that is still producing water exists at the bottom of East Fork Navajo Canyon, next to Chapin Mesa. This spring would likely have been a distant secondary water supply to which those who lived on this mesa had to trek when Sagebrush Reservoir was dry.

3.4. Box Elder Reservoir

The final reservoir investigated by the WWE/WPI team was Box Elder Reservoir, a valley-bottom water storage site in Prater Canyon about 1.2 km west of Morefield Canyon. Box Elder Reservoir strongly resembles Morefield Reservoir as they are both on an east–west latitude at the bottom of canyons, they were similarly constructed, and they both collected groundwater and runoff from nearby agricultural fields. Box Elder Reservoir was unknown to MVNP staff until the Bircher Fire of 2000 consumed the brush and trees that covered the mound left by the reservoir (Figure 16). A park ranger noticed the mound and saw it had many of the hallmarks of Morefield Reservoir. The WWE/WPI team conducted auguring and other studies that found water-borne sediment, layers of dredging evidence, and potsherds, confirming the Box Elder site had been a reservoir.

WWE/WPI team research suggests that the Box Elder Reservoir was initiated by Pueblo I people in 800 CE or earlier. Based on precipitation data, the reservoir would have stored water about five times per year. Eventually the reservoir rose so high in elevation due to sediment accumulation and dredging that a diversion canal and increased maintenance became necessary. Team member Gregory Hobbs searched the bed of the Prater Canyon Creek for remnants of an ancient canal diversion that would have fed the reservoir and found none. However, while tracking the likely route of the canal between the creek bed and the reservoir area, he found an erosional remnant of a canal with typical canal liner stones. Interspersed with the stones were clay potsherds that Dr. Breternitz identified as Pueblo I era. This helped verify our hypotheses that this was the ancient diversion canal that carried water to Box Elder Reservoir.

By about 950 CE, the sediment from the canal caused the Box Elder Reservoir to reach six meters in elevation, become unmanageable, and no longer be viable for water storage. Park archeologists and the WWE/WPI team surmised that Morefield Canyon residents transferred technical reservoir-building knowledge to the residents of Prater Canyon around 800 CE (Figure 17) based on the many similarities of the two storage structures, the earlier evolution of the Morefield Reservoir, and the proximity of the two sites to each other.

Figure 17 shows that both reservoir sites had nearby springs that could also have been used for water supply, although springs were far more abundant in and near Morefield Canyon than Prater Canyon. This is congruent with an estimate by park archeologists that Prater Canyon had about 60% of the population of Morefield Canyon.

3.5. Mug House Cistern

After the reservoir research, the WWE/WPI team obtained a permit to study a water collection cistern located near the Mug House Ruin (Figure 18). This cliff dwelling on Wetherill Mesa was home to about 80–100 Pueblo III people during the 13th century. As conditions became drier and competition for resources surged, the ancestral Pueblo people began to develop cliff dwellings in locations that were more protected and defensible than the pit houses and kivas. Mug House was abandoned abruptly around 1280 CE, likely due to encroaching raiders [8].

The Mug House cistern (Figure 19) is a 15,000-L water storage vessel at the base of a 26-m-high cliff. At the top of the cliff, there is a deep notch that serves as a collection point for a six-acre forested drainage basin. A WWE/WPI team of about 20 professionals performed topographical surveys, inspected the drainage basin, analyzed the cistern structure, collected soil samples for soil gradation testing, performed hydrological analyses, and collected pollen samples to determine pre-historic vegetation.

The team conducted eight flow tests, discharging water at different flow rates at the collection notch to simulate ancient runoff as a rappeler recorded video of the hydraulic jet formed by the notch (Figure 20). The team learned that the hydraulics of the notch were functional and delivered water to the cistern at a range of flows. The notch was formed, either by nature or through manipulation, to jet the water away from the cliff wall so that little water was wasted during the drop.

The cistern is roughly rectangular with dimensions of 2.5 by 7 m, with the longer axis of the cistern running parallel to the cliff face. The maximum depth for water storage before overtopping is one meter. On the west side of the cistern, the ancient engineers incorporated an overflow port allowing excess water to safely exit the structure. At the bottom of the north end of the cistern are two large sandstone splash blocks. Splash blocks minimize erosion due to impact from the high-velocity water falling from the cliff above, keeping the cistern floor intact and reducing turbidity levels that affect water quality. The blocks were also well-placed to be used as a jug-filling platform so that a water gatherer could avoid stepping onto the cistern bottom and raising clouds of sediment when the water level was low.

The ancestral Puebloans capitalized on 230 square meters of bare sandstone rock above the notch that shed rapid and immediate runoff when it rained. The team noted that the ancient people of Mesa Verde were likely aware of the need to control channel erosion. In the drainage basin above the Mug House cistern, the ancestral Pueblo people built sandstone block erosion control checks in two natural channels that convey water from east to west to the cliff edge. These ancient structures helped stabilize the channels by preventing downward channel cutting.

Hydrological analyses indicate that the Mug House cistern, under average conditions, would fill to capacity (15,000 L) from a single rainfall event about once each year. About 3400 L would have been generated four times per year, and twice per year, it would receive 4500 L. About 40 times per year, the cistern would receive 760 L of water, or more, per rainfall event. In a typical year, the cistern would have received approximately 380,000 L of water supply for useful storage, which could have supplied most of the domestic water needs of the Mug House community during approximately the April through October period. The team judged that the Mug House water gatherers would have had to trek much farther to a local spring for water during other times of the year [9].

3.6. Goodman Point Spring

The WWE/WPI team’s final ancestral Pueblo water supply study site was Goodman Point, the one site described in this paper that is part of Hovenweep National Monument rather than MVNP. Goodman Point, approximately 18 km northwest of Cortez, Colorado, was occupied during roughly the same period as Mug House. The settlement was a walled community of about 500 to 800 people who had earlier lived in scattered family farmsteads but gathered at Goodman Point, likely for increased security, around 1260 CE. Like Mug House, Goodman Point was abandoned abruptly around 1280 CE [10].

Team studies focused on Juárez Spring (Figure 21, which was within the walled settlement and would have been the main source of water for the Goodman Point community. The team also researched other water sources that would have existed at the time: an ancient reservoir, now called Goodman Lake, and Mona Spring. Both Goodman Lake and Mona Spring were about one kilometer south-southwest of the Goodman Point Pueblo. Juárez Spring water issues from the Dakota Sandstone at two distinct points designated “Juárez Spring No. 1” and “Juárez Spring No. 2”. The distance between the two springs is about 24 m. WWE/WPI studied geology, soils, spring location, water quality, and ground infiltration rates to assess the water supply provided by Juárez Spring. The team greatly benefitted from prior archeological studies performed at Goodman Point by the Crow Canyon Archaeological Center.

Juárez Springs Nos 1 and 2 were developed on the sides of the slopes where the water issued from the bedrock. These locations would allow for the direct filling of jars. Below each spring outlet were pools that collected spring water. Because Juárez Spring has a low flow rate of about 10,600 L per day, it would have been necessary for Goodman Point residents to use both water directly from the spring and the pooled water. There would have been enough water to maintain the turkeys that were a major food source, but not for irrigation of crops. Luckily, dryland farming was, and still is, possible in the Four Corners area [11].

Mona Spring issues from a sandstone outcrop near the bottom of a shallow gully. Here, the likely headworks was a gully-bottom pool. Goodman Lake, with a sporadic water supply, provided stored water from direct rainfall-runoff (Figure 22). Stepping stones in the lake bottom would have allowed access to the pooled water for the filling of jars. The ancestral Pueblo people could have walked the kilometer to Goodman Lake or Mona Spring to fill their jars if necessary, helping ensure an adequate supply.

The team determined that the quality of the Juárez Spring water was good for its purposes, even with rather high total dissolved solids levels. The water was essentially free of harmful organisms, such as bacteria. However, the pooled water adjacent to the springs and in Goodman Lake was susceptible to contamination [12].

4. Conclusions

The six ancestral Pueblo community water supply sources investigated by the WWE/WPI team provide much insight into these people and their circumstances. Through these artifacts, the ancestral Pueblo people demonstrated an impressive application of water harvesting knowledge and skill. They showed good hydrological understanding through their water storage and transportation activities, and also displayed organizational abilities through their collaborative creation and maintenance of reservoirs over many generations. The climate of this period was harsh, but the ancestral Pueblo people pulled together to create effective community systems that allowed them to survive with limited resources.