Soil-Geomorphic and Soil Physical Change

Physical soil management changes in soil surfaces and topography in arid land agriculture are primarily intended to increase and control water supply to crops, moderate temperature, create or promote crop physical stability, increase soil volume for crop rooting (and so increase water and nutrient uptake capacities), and to protect soils from excessive erosion.

Terracing is among the most visible and intensive kinds of management in ancient agricultural landscapes. Terracing involves segmenting slopes into topographic steps and has been used extensively in sloping terrain to create stable, nearly level fields and to manage water (Denevan 2001; Donkin 1979; Doolittle 2000; Sandor 2006; Sandor and Homburg 2016). The stepped topography resulting from terrace wall construction and sediment filling is generally characterized by reduced field slope angle and length. Along with creating a stable topographic base for crops, functions of terracing include soil retention and erosion control, soil building by construction filling or sedimentation, microclimatic modification, and water control ranging from runoff management to irrigation and drainage. Intensively terraced irrigated landscapes are especially prevalent in the Andes and parts of Mexico. The Colca Valley, Peru, is an example of monumental bench terracing and irrigation, in which much of the land below 4000 m elevation has been transformed into a terraced landscape (Denevan 2001; Sandor and Eash 1995; Sandor et al. 2006; Treacy 1989; Figure 2). Terrace agriculture in the Colca Valley has been practiced for over 1500 years and continues today. Terracing in the Andes has been practiced for at least 4000 years (Denevan 2001) and there is evidence for irrigation as early as 6700 yrs BP (Dillehay et al. 2005). Intricately constructed canals woven as passages through flights of durable stone terraces constitute a monumental and enduring transformation of soils and landscape in the Colca Valley and other terrace systems in the Andes and Mesoamerica.

Figure 2.

Terraced landscape and soil in the Colca Valley, Peru (upper photo). Thick anthropogenic, organic matter-rich soil exposed where agricultural terrace wall partially removed by farmers for repair (lower photo).

Beneath terraced surfaces, soils have been altered to improve production (Figure 2). One of the most distinctive morphological features of terraced soils is increased surface horizon (A horizon) thickness resulting directly from accumulation of material upslope of terrace walls either by hand-filling or sedimentation. The amount of soil thickening within each terrace varies because of the wedge-shaped geometry common in terraced soil fills, wherein fills are thickest nearest the terraced wall and decrease upslope (Sandor 2006). In the case of the Colca Valley, A horizon thickening mostly ranges from 0.3 to 1 m and the range reported elsewhere is a few centimeters to several meters. Other positive physical soil changes resulting from management are more stable granular soil structure, increased macroporosity, and reduced bulk density (Sandor and Eash 1991, 1995). These structural changes are imparted by careful management practices, such as long-term organic matter additions and tillage practices that minimize soil disturbance. Anthropogenic changes in the fine earth texture of soils were not observed in the Colca Valley, but gravels were emplaced in soils just upslope of terrace walls to aid water drainage and prevent excess pressure on terrace walls. In prehispanic terraces in Chile, rock fragments were removed from upper soil horizons (Sandor et al. 2015). All the physical soil changes described enhance crop production by increasing topsoil volume and tilth that can be exploited by roots, thereby increasing and facilitating water and nutrient uptake. Specific alteration of soil physical properties for management in terraced soils in the Americas has also been reported by Field (1966), Keeley (1985, 1988), and Sampietro Vattuone et al. (2011, 2014). Some studies of ancient agricultural terraces have revealed complex stratigraphy, with multiple soils and paleosols, indicating multiple use and reconfiguration of terraces during different time periods (Borejsza et al. 2008; Branch et al. 2007; Pérez Rodríguez and Anderson 2013; Sánchez-Pérez et al. 2013; Zaro and Umire Alvarez 2005).

Another form of intensive soil change in ancient agriculture is the development of greatly thickened soils through long-term accumulation of suspended sediment in fields from irrigation canals. This sedimentation process over centuries results in the formation of thick anthropogenic “irragric” horizons, whose uniform textures are commonly rich in fine particles (fine sands, silts, or clays). While irragric horizons cannot always be unequivocally attributed to deliberate management, they commonly improve soils for agriculture by increasing topsoil volume, replenishing soil fertility, and reducing soil salinity. Awareness of the benefits of moderate sedimentation in traditional agricultural fields in renewing and improving soils and crop production indicates that, in many cases, this is intentional management (Doolittle 2000; Nabhan and Sheridan 1977; Russell 1908; Williams 1981). However, a balance between the benefits of slight to moderate sedimentation and the detrimental effects of excess sedimentation was not easy to achieve (e.g., Dart 1986; Woodson et al. 2015). An example of an irragric soil is that developed in fields in the Gila River Valley in the Phoenix Basin from about AD 450–1450 by the Hohokam, whose remarkable canal irrigation systems—the largest in the Americas north of Peru—are renowned for their monumental scale and skilled engineering (Doolittle 2000; Fish and Fish 2007). In this case, a silt-rich irragric horizon about 77 cm thick with favorable properties for agriculture buries a natural clay-enriched argillic horizon on a Pleistocene stream terrace (Woodson et al. 2015; Figure 3). Other examples of irragric soils in ancient fields in the American Southwest have been reported (Hall et al. 2013; Homburg 2015; Huckleberry 1992; Means 1901; Schaafsma and Briggs 2007). Along the south coast of Peru, irragric soils dating to more than 3500 yrs BP and up to 4 m thick were documented by Hesse and Baade (2009).

Figure 3.

Prehistoric irragric soil and chemical properties related to salt management. Snaketown canal-field system, middle Gila River Valley, Arizona. Photo scale is divided into 10 cm bands. In the table, numbers shown are means of soil samples from inside, adjacent to, and outside of the prehistoric irrigated field area. Asterisks indicate statistically significant differences among locations with the Kruskal–Wallis test at the following probability levels: *P < 0.05, ***P < 0.001. See Woodson et al. (2015) for more information.

Subtle, yet widespread, cases of landscapes and soils physically altered by prehistoric agricultural management are also evident. An example is the extensive and diverse dryland agricultural fields in the American Southwest and northern Mexico located in valley margin and upland environments (Adams 2004; Doolittle 2000; Herold 1970; Sandor and Homburg 2015). Some of these fields are identified from terracing and other rock configurations used to create and manage the fields. Examples of these landscape modifications using rock alignments are terraces, grids, piles, and lithic mulch fields (Figure 4). However, these structures were not always used, so it is likely that many early dryland fields remain unrecognized.

Figure 4.

Prehistoric dryland agricultural fields in the American Southwest managed with rock configurations. A - Lithic mulch field, Chama River Valley, northern New Mexico; B - Grid fields, Gila River Valley near Safford, eastern Arizona. Average grid size is about 22 m² (Doolittle and Neely 2004). Aerial photo by James A. Neely; C - Rock pile field and agave near Casas Grandes, Chihuahua, Mexico; D - Terraces for runoff agriculture, Sapillo Valley (near Mimbres Valley), southwestern New Mexico.

Because water is the most limiting resource in arid land agriculture, a key function of rock configuration fields relates to water conservation. For example, rock piles helped retain soil moisture for agave and other desert crops by reducing evaporation (Fish and Fish 2014; Homburg and Sandor 2011). Lithic mulch fields operated on a similar basis and another likely primary purpose of lithic mulch in the extensive fields of the northern Rio Grande region in New Mexico was to moderate temperature—especially to promote warmer spring temperatures in the soil during the early growing season at the higher latitudes and elevations of the northern Southwest (Anscheutz 1998; Lightfoot 1996; White et al. 1998).

The most widespread kind of dryland agriculture in the Southwest and other arid regions is runoff farming. An important achievement of runoff agriculture is its ability to supply water and nutrient needs to crops without conventional irrigation or fertilization. This is done by connecting fields to hydrologic and ecosystem processes in watersheds. Fields are carefully placed and managed on certain landscape positions, such as alluvial fans, to intercept storm runoff and associated sediment and organic debris transported from adjoining uplands. Within fields, terracing or analogous management boosts water retention by reducing slope angle and length and by encouraging sedimentation. The sediment becomes a thicker topsoil with a desirable loamy to sandy texture that increases water infiltration, rooting volume, and thus water-holding capacity for plant use. In some cases, erosion of soils in watersheds was used as a management tool to encourage sediment deposition in fields downslope (Doolittle 2000). In-depth knowledge among traditional farmers about soil, geomorphic, hydrological, and ecological processes involved in runoff agriculture in the American Southwest and Mexico has been recorded in some studies (e.g., Nabhan 2013; Nabhan and Sheridan 1977; Pawluk 1995; Sandor et al. 2007).

Soil Chemical and Biological Change

Management of soil chemical and biological properties in arid land agriculture is primarily intended to enhance or replenish nutrients removed from soil by crops and to protect soils from accumulating excess salt and sodium that harm crops.

Chemical and biological changes in ancient agricultural soils resulting from management have been inferred in a few studies (Table 1). In the Colca Valley, Peru, higher levels of organic carbon, nitrogen, and phosphorus were found in agricultural soils compared with uncultivated soils (Sandor and Eash 1991, 1995; Figure 5). Extraordinary levels of phosphorus were found in some subsurface horizons of agricultural soils, indicating additions of P-rich fertilizer materials and significant downward movement of phosphorus over centuries. Higher soil enzyme activity associated with nitrogen and phosphorus metabolism was also detected (Dick et al. 1994). These soil changes were attributed to the long-standing Andean practice of fertilization with animal manure and other materials, including sea-bird guano. Increases in soil nutrients in traditional Andean agricultural soils fertilized with manure were also measured by Winterhalder et al. (1974). These practices and others—such as crop rotation, fallowing, diverse intercropping, and use of nitrogen-fixing legumes—have been instrumental in conserving and improving soils. These show long-term soil care resulting from deliberate fertilization and organic matter management. The substantial knowledge underlying this impressive management is exemplified in the traditional soil classification system in the Colca Valley (Sandor and Furbee 1996; Sandor et al. 2006).

Figure 5.

Examples of soil chemical change in ancient agricultural soils.

Upper graph: comparison of organic carbon, total nitrogen, and total phosphorus in terraced agricultural and uncultivated upper A horizons in the Colca Valley, Peru. Bars are means (lines within bars show ± 1 standard deviation) from each sample group. Mean differences are statistically significant minimally at the 0.05 probability level in all cases except between current and abandoned agricultural soils for organic carbon and total nitrogen. See Sandor and Eash (1991, 1995) for more information.

Lower graph: comparison of total nitrogen in uncultivated and Mimbres prehistoric agricultural soils. Bars are means (lines within bars show ± 1 standard deviation) from prehistoric cultivated soil, uncultivated soil, and recent runoff sediment sample groups. ** = significant overall differences among sample groups at the 0.01 probability level. Different letters indicate individual mean differences at the 0.05 probability level. Refer to Sandor et al. (1986, 1990) for more information.

In other arid regions of the Americas, such as the North American Southwest, where there is little evidence for direct application of fertilizer materials, higher levels of nutrients have not been consistently detected in ancient fields. However, there is soil and ethnographic evidence of nutrient replenishment by watershed hydrologic processes in agricultural systems that use runoff. Runoff management not only increases water supply for crops, it also creates new soil and replenishes fertility as runoff delivers fresh nutrients. Studies of watersheds and linked fields at Zuni documented soil build-up and maintenance of soil nutrients (Homburg et al. 2005; Norton et al. 2007; Sandor et al. 2007). Studies in ethnoecology and ethnopedology indicate that traditional farmers in the Southwest and Mexico are well-aware that runoff inputs maintain and enrich soils (Nabhan 2013; Pawluk 1995; Sandor et al. 2007; Wilken 1987). Agricultural experiments with runoff and monitoring runoff materials in agricultural watersheds have provided data about the processes of nutrient decomposition and availability to crops, as well as the effects of nutrients in runoff on crop productivity (Hubbell and Gardner 1950; Norton et al. 2007). A few studies of soil biological activity point to the importance of nitrogen-fixing organisms and plants, mycorrhizal fungi, and biological crusts in maintaining soil fertility in traditional agricultural systems (Berry 1995; Havener 1999; Norton et al. 2007). Inputs of nutrients to soils through eolian dust have also been found (Nakase et al. 2014) and recognition of wind transport of sediment in replenishing soil has been reported ethnographically (Cushing 1920; Ferguson 1985; Pawluk 1995).

Controlling salt is an aspect of soil chemical property management especially important in arid lands. Salt and sodium accumulation and associated high pH are a major hazard in past and present irrigation agriculture that can seriously damage land (Brady and Weil 2008). Crop tolerance to salt and sodium is variable: among major crops of the Americas, beans (Phaseolus spp.) are highly sensitive, cotton (Gossypium hirsutum) is more tolerant, and maize (Zea mays) is intermediate. Damaging effects of salt in early agriculture have been inferred in other world regions, such as Mesopotamia (commonly accepted as true, though some controversy exists), and salinization has been implicated as a factor in the demise of the Hohokam in the American Southwest (Diamond 2005; Reisner 1986). However, these assertions are not based on direct soil evidence. Evidence from a recent study of Hohokam irrigated soils from the Snaketown canal-field system along the middle Gila River Valley indicates that salt, sodium, and pH are actually lower in the irrigated soils compared with adjacent similar land outside the system, although there are salt levels that need to be managed (Woodson et al. 2015; Figure 3). Lower salt and sodium in the Hohokam fields can be attributed to two factors: 1) the younger age of the anthropogenic soil so that there was less time for salt accumulation and 2) the leaching of salts by applied irrigation water. While it is not certain that this is due to deliberate management, indirect evidence for management of salt is: 1) that the Snaketown irrigation system functioned for a long time (7–10 centuries) and 2) an ethnographic record of careful salt management among historic Indian farmers in the area. More research is needed, but the longevity of the Hohokam irrigation systems on the order of a millennium suggests that they were able to manage salinity problems (Ackerly 1988).

Unintended Soil-Geomorphic, Ecological, and Soil Physical Change

Unintended landscape and physical soil degradation from agriculture documented in ancient fields in arid lands mainly involves accelerated erosion (by wind, water, and mass movement) that can remove soil, particularly valuable topsoils. Other deleterious impacts recognized in ancient agricultural fields are compaction and soil structural degradation that restricts root growth and reduces water and nutrients available to plants. Accelerated erosion (including that following abandonment of ancient agricultural terraces) and other physical soil degradation, such as soil compaction, have been reported in several studies in Mexico (Borejsza et al. 2008; McAuliffe et al. 2001; McClung de Tapia 2012; Sánchez-Pérez et al. 2013), Peru (Goodman-Elgar 2008; Inbar and Llerena 2000; Zaro et al. 2008), and the American Southwest (Kruse-Peeples et al. 2009).

One case of such degradation is in prehistoric upland agricultural fields in the Mimbres region of New Mexico (Sandor and Homburg 2015; Sandor et al. 1986, 1990). These fields mainly date to the Mimbres Classic period (AD 1000–1130), based on archaeological evidence of association with nearby small pueblos and ceramic sherds within terraced soils. This dryland agriculture used a series of low rock terraces to capture runoff from summer monsoon storms. Testing for soil change was possible because intact fields could be identified, control areas were available, and post-agriculture land use was relatively minimal. Soils were sampled in transects across agricultural and unfarmed areas with similar geomorphic settings. Using the comparative approach, major soil-landscape changes from agriculture were inferred, namely long-term soil and ecosystem degradation. Widespread gully and sheet erosion was observed in the agricultural area but not within the unfarmed reference areas. Evidence that accelerated erosion began during prehistoric agriculture is that terraces are breached by gullies, showing that terracing came first, and attempts by prehistoric farmers to repair the damage are shown by rock dams built within gullies. Long-term anthropogenic ecosystem change involving reduced grass cover contributed to the accelerated erosion, which continues to the present. Besides erosion, degradation of soil physical properties within fields is also apparent. Upper soil horizons have suffered structural degradation and compaction, and lighter soil color reflects lower organic matter levels.

Unintended Soil Chemical and Biological Change

Degradation of soil fertility and biochemical properties from agriculture documented in ancient fields of arid lands mainly involves decreased levels of organic matter and nutrients, increased levels of salt and sodium, and accompanying changes in soil biology. In the prehistoric Mimbres case, biochemical degradation is expressed in the form of decreased organic matter and key nutrients nitrogen and phosphorus in upper soil horizons (Sandor and Homburg 2015; Sandor et al. 1986, 1990; Figure 5). Decline in soil organic matter is linked to the soil structural degradation and compaction. A controlled greenhouse experiment comparing prehistoric agricultural and unfarmed reference soils showed that maize growth and nitrogen content were reduced in the degraded agricultural soil (Sandor and Gersper 1988). Lack of soil recovery centuries after cultivation ended illustrates how degradation can persist for a long time in environments sensitive to disturbance, in this case about a millennium.

Other cases of soil organic matter and nutrient loss in ancient agricultural soils have been reported in the American Southwest (Berlin et al. 1977, 1990; Edwards 2002; Sullivan 2000) and the Andes (Goodman-Elgar 2008; Sampietro Vattuone et al. 2014). Anthropogenic salinization has been reported in prehistoric fields in the Grand Canyon by Davis et al. (2000).

A current study in the Atacama Desert of Chile (Sandor et al. 2015) provides an example where unintentional soil change is not necessarily all bad. In this hyper-arid climate, agriculture is only possible with irrigation, so ancient terraced agricultural soils can be compared with nearby natural soils upslope from the highest canal. One major soil change is an influx of calcium carbonate by long-term irrigation with calcareous spring water. The downside for crops is that the resulting higher pH lowers availability of plant nutrients, such as phosphorus. But a possible benefit of this carbonate for crops is increased soil water-holding capacity (Duniway et al. 2010; Georgen et al. 1991).