476 SSSAJ: Volume 71: Number 2 • March–April 2007

Soil Sci. Soc. Am. J. 71:476–483doi:10.2136/sssaj2006.0019Received 13 Jan. 2006.*Corresponding author (jnorton4@uwyo.edu).© Soil Science Society of America677 S. Segoe Rd. Madison WI 53711 USA

Sustained productivity of ancient agricultural soils on alluvial fans of the Zuni Indian Reservation exemplifi es the importance of

hillslope processes that link uplands to alluvium-derived soils (Bull, 1997; Peterson et al., 2001). Many ancient agricultural fi elds on alluvial fans in North America’s Colorado Plateau continue to be farmed by Native Americans, including the Zuni of western New Mexico (Damp et al., 2002; Muenchrath et al., 2002; Homburg et al., 2005; Sandor et al., 2007). Zuni farmers do not use added fertilizers in traditional runoff agriculture but recognize the role of upland hillslopes in sustaining production of corn (Zea mays L.) and other crops in the semiarid environment. They work to enhance processes that link hillslopes to their fi elds by preventing channel incision and diverting runoff (Cushing, 1920; Norton et al., 2002). Farmers interviewed by Pawluk (1995) defi ned sediment as “good rich soil… from up there; rains bring the soil down… where we get the nice fertilizer from.” They described upland hillslopes as contrib-uting “all kinds of fertilizers and water” to “the rich place a little ways out from the wash” (Norton, 2000). This ancient agroecosystem creates an excellent setting for investigating contributions by upland hillslopes to nutrient cycling and retention in alluvium-derived soils of headwater ephemeral streams, which are often the most biologi-cally diverse and productive landforms in semiarid landscapes.

Hillslope runoff and sediment transport processes have been described by many researchers (Leopold et al., 1966; Schumm and Mosley, 1973; Selby, 1993; Bull, 1997). The unique terrain of the Colorado Plateau creates lithologically

segmented hillslopes that control the distribution of vegeta-tion and soils (Norton et al., 2003), as well as the quantity and composition of runoff and sediments. Our objective was to describe contributions of upland hillslopes to downslope allu-vial fans as infl uenced by slope position, cover type, and rainfall characteristics on the Zuni Indian Reservation in New Mexico. We focused on hillslopes in headwater drainages where archae-ological evidence shows that downslope alluvial fans have been farmed by Native Americans for >1000 yr.

Previous work established that organic C and total N and P contents along summit-to-toeslope transects follow parabolic trends that peak in soils of wooded BS on the hillslopes of our study area. Mineral N and P contents in the same soils increase linearly from summits to toeslopes (Norton et al., 2003). This suggests that hillslope processes mix and decompose organic materials as they are transported downslope toward agricul-tural fi elds. In this study, we investigated how slope position and cover type infl uence the yield and composition of hillslope sediments and organic materials. Additionally, we investigated the relationships between sediment composition and rainfall parameters in this distinctive, summer-rainfall-driven system.

MATERIALS AND METHODSIn this three-part study, we analyzed the quantity and composi-

tion of summer hillslope runoff as infl uenced by (i) slope position, (ii) cover type, and (iii) rainfall characteristics. We collected data in three small watersheds above long-term runoff agricultural fi elds on the Zuni Indian Reservation, New Mexico, in the southeastern part of the Colorado Plateau physiographic province (Fig. 1). The reservation lies at 1800- to 2400-m elevation and receives an average of 300 mm of pre-cipitation annually, much of which often comes during thunderstorms in July, August, and September (Fig. 2). In Zuni runoff agriculture, farmers rely on deep soils on alluvial fans to store winter precipitation

Runoff and Sediments from Hillslope Soils within a Native American AgroecosystemJ. B. Norton*Dep. of Renewable ResourcesUniv. of WyomingLaramie, WY 82071-3354

J. A. SandorDep. of AgronomyIowa State Univ.Ames, IA 50011-1010

C. S. WhiteDep. of BiologyUniv. of New MexicoAlbuquerque, NM 87131

Farmers of the Zuni Indian Reservation in New Mexico rely on materials transported from upper watersheds to maintain productivity of some of the oldest agricultural fi elds in North America. This study determined runoff and sediment production from hillslopes as functions of slope position, soil cover, and rainfall characteristics. Runoff was collected from plots in summit or shoulder [SU], backslope [BS], and footslope [FS] positions and in bare soil, microbiotic crust, oak, juniper, pinyon, and grass cover. In rainfall events that generated runoff, sediment C and N concentrations decreased with total sediment yield, but C/N ratios increased. Carbon/nitrogen ratios were generally lower in sediments from early season events. Backslope plots yielded the most sediment (125 g m−2) and the most organic C (2.2 g m−2), total N (116 mg m−2), and total P (34.2 mg m−2), but those from SU plots had the highest concentrations (e.g., 31 g C kg−1 compared with 23 g C kg−1 at BS and FS). Bare soil and microbiotic crust plots yielded the most sediments (441 g m−2) and grass the least (107 g m−2). Though variable, oak plots yielded more C (32.6 g m−2) and N (1.5 g m−2) than others, but bare soil and microbiotic crust plots yielded the most P (62.7 and 54.4 mg m−2, respectively). Slope position, cover type, and rainfall characteristics interact to infl uence movement and processing of materials responsible for agriculturally, ecologically, and hydrologically important alluvium-derived soils in this semiarid agroecosystem.

Abbreviations: BS, backslope; FS, footslope; OM, organic material; P, phosphorus; SOM, soil organic matter; SU, summit or shoulder.

FOR

EST,

RA

NG

E &

WIL

DLA

ND

SO

IL Published online March 12, 2007

SSSAJ: Volume 71: Number 2• March –April 2007 477

(typically snow and low-intensity rainfall) for early crop growth and on summer precipita-tion for crop maturity and grain production.

The Sanchez watershed covers 68 ha, the Laate watershed 7 ha, and the Weekoty watershed 125 ha. Each of the watersheds lies in the eastern half of the Zuni Reservation at about 2070-m elevation (Fig. 3). Distribution of soils and vegetation corresponds to slope positions (described by Norton et al., 2003). Broad, level summit and shoulder positions, with 0 to 20% slopes, have shallow, sandy Entisols covered by open pinyon(Pinus edu-lis Engelm.)–juniper (Juniperus spp.)–oak (Quercus gambelii Nutt.) woodlands with appreciable ponderosa pine (Pinus ponderosa P. & C. Lawson), shrubs such as wavy-leaf oak [Quercus × pauciloba Rydb. (pro sp.) (gambelii × turbinella)], mountain mahogany (Cercocarpus montanus Raf.), fragrant ash (Fraxinus cuspidata Torr.), and antelope bit-terbrush [Purshia tridentata (Pursh) DC.] and Douglas-fi r [Pseudotsuga menziesii (Mirbel) Franco] in protected areas. Microbiotic crusts and native grasses (e.g., Stipa spp., blue grama [Bouteloua gracilis (Willd. ex Kunthi) Lag. ex Griffi ths], squirreltail (Elymus elymoides [Raf.] Swezey ssp. elymoides), and mutton grass [Poa fendleriana (Steud.) Vasey]) cover open areas on summit and shoulder positions.

Backslopes, with 20 to 70% slopes, and FS, with 2 to 25% slopes, are covered by more dense pinyon–juniper–oak vegetation with very little ground cover or microbiotic crusts in interspaces. Backslope soils formed in loamy colluvium over siltstone residuum and grade from shallow Entisols on upper BS to deep, well-developed Alfi sols on FS. Toeslopes, with 0 to 10% slopes, are covered by big sagebrush (Artemisia tridentata Nutt.), western wheatgrass [Pascopyrum smithii (Rydb.) A. Löve], and blue grama with components of rabbitbrush [Chrysothamnus nauseosus (Pallas ex Pursh) Britt] and weedy herba-ceous vegetation in wash areas.

Runoff and Sediment by Slope PositionWe collected and analyzed summer storm runoff (approximately 1

May to 31 August) from 14 sediment traps below 20-m2 bounded plots (2.5 by 8 m) (Williams and Buckhouse, 1991; Gellis, 1998) at the three study sites (four mesa top, six BS, and four FS plots; see Fig. 3 and Table 1). Data were collected from plots at the Sanchez site (two BS and two FS plots), the Laate site (two SU and two BS plots), and the Weekoty site (two SU plots) during the summer of 1996 and from the Weekoty site (two SU, two BS, and two FS plots) during the summers of 1997 and 1998 (Table 1). Yield–landscape position relationships were consistent among the sites and data-collection periods. The plots were installed at locations with vegetation and soils representative of each slope position. Samples were captured in 19-L buckets from which 4-L, thoroughly mixed subsamples were collected after each runoff event. The depth of runoff trapped in col-lection buckets was measured for calculation of runoff volume by event. Subsamples were allowed to settle in refrigerators at 4°C and then decanted. Sediment samples were air dried and stored for lab analyses. The super-natant was preserved with a dilute solution of phenylmercuric acetate and

stored for solute analysis. Rainfall was recorded from rain gauges at each runoff plot and at a tipping bucket rain gauge equipped with a CRX-20 data logger (Campbell Scientifi c Equipment, Logan, UT) located in the Weekoty watershed.

Runoff and Sediment by Soil Cover TypeWe installed 12 1-m2 bounded plots (2 by 0.5 m) with sediment

traps, two plots in each of six cover types, at the Weekoty site (bare soil, grass, pinyon pine, juniper, oak, and microbiotic soil crust; see Fig. 3 for approximate locations). Plots were selected to represent pure stands of each prevalent cover type as determined in Norton et al. (2003). Ground beneath the trees was covered by litter of that species without appreciable amounts of other vegetation, which is typical for this semiarid pinyon–juniper woodland. The plots were located on FS

Fig. 1. Study area location and physiographic provinces of the Southwest.

Fig. 2. Long-term average monthly precipitation (1949–2005) for Zuni, NM (from the Western Regional Climate Center (2006).

478 SSSAJ: Volume 71: Number 2 • March –April 2007

with 7 to 12% slopes. The sediment traps were monitored during the summers of 1997 and 1998. Sediments and runoff were measured and treated the same as for the 20-m2 slope position plots.

All sediment samples were oven dried at 105°C. Particle-size dis-tribution was determined using the sieve and pipette method (Gee and Bauder, 1986). Total C and N concentrations were determined using a Fissions EA1100 dry combustion CNSHO analyzer (Fissions Instruments, Milano, Italy). Inorganic C was determined by coulomb-meter in a subset of samples and found to be insignifi cant relative to total C values. Total P concentrations were determined by alkaline oxidation (Dick and Tabatabai, 1977). Available-P concentrations were measured by the Olsen extraction method (Olsen and Sommers, 1982). Runoff samples were analyzed for cation concentration by

atomic absorption spectrophotometry and anion concentration with a Technicon AutoAnalyzer (Technicon, Tarrytown, NY).

Data AnalysisThe yield and composition of runoff water and sediments were ana-

lyzed by slope position, cover type, and several rainfall parameters. Also, yields were estimated for whole watersheds. For analyses by slope position and cover type, we analyzed runoff and sediment data by average summer (May–August) yield and average concentration during the study period.

For runoff and sediment yield, the experimental unit averaged for analysis of means (basis for n) was the total amount of a given constitu-ent collected during each summer at each runoff plot. This amounted to a total of eight plot-years at the SU, eight at the BS, and six at the FS slope

position runoff plots (see Table 1) and four plot-years at each cover-type runoff plot. Although we also ana-lyzed yields by event, only results for total summer yields of each constituent are presented here because they represent gross movement of materials toward agricultural fi elds.

Concentration was analyzed by runoff event. The experimental unit averaged for analysis of means (basis for n) was the concentration of a given constitu-ent collected after each runoff event that yielded sedi-ment at each plot during the entire study period. This amounted to 34 events at the SU, 39 at the BS, and

Table 1. Number of runoff plots monitored at each study site, growing season, and slope position†.

Watershed1996 1997 1998 TOTAL‡

SU BS FS SU BS FS SU BS FS SU BS FS

Sanchez – 2 2 – – – – – – - 2 2

Laate 2 2 – – – – – – – 2 2 -

Weekoty 2 – – 2 2 2 2 2 2 6 4 4

Total yr−1 4 4 2 2 2 2 2 2 2 8 8 6

† SU = summit; BS = backslope; FS = footslope.

‡ Total by slope position is the basis for n in Table 4.

Fig. 3. Topography, slope positions, and runoff plot locations within the three watershed study sites on the Zuni Indian Reservation (modifi ed from Norton et al., 2003).

SSSAJ: Volume 71: Number 2• March –April 2007 479

31 at the FS runoff plots. At the cover-type runoff plots, bare soil plots yielded runoff 27 times during the two summers, microbiotic crust 30 times, oak cover 26 times, juniper cover 25 times, pinyon cover 22 times, and grass cover 24 times.

Means were compared by calculating least signifi cant differences with the GLM procedure (SAS Institute, Cary, NC). To ensure normal distri-bution and equivalent variance, particle-size distribution data (presented as percentages) were arcsine transformed before statistical analyses.

We examined sediment yield and composition from hillslopes in the Weekoty watershed as functions of several characteristics of rainfall mea-sured at the tipping-bucket rain gauge. Maximum rainfall intensity, rainfall duration, duration of maximum intensity, time from start of event until maximum intensity, and length of rainless period before start of event (Ferreira, 1990) were calculated for each event based on a 15-min gap to distinguish events. Each parameter was correlated against runoff and sedi-ment properties from 1997 and 1998 at the Weekoty watershed by regres-sion analysis (trend line function in Microsoft Excel). Total sediment yield from each event is presented here as a surrogate for rainfall parameters because it integrates the many variables that affect relationships between rainfall and runoff and reveals relationships between sediment composition and storm intensity.

To estimate yields of runoff, sediment, and selected constituents by whole watershed, we extrapolated the runoff and sediment data from slope position and cover-type plots to the whole watersheds by integrating with spatial distribution data presented in Norton et al. (2003).

RESULTS AND DISCUSSIONRainfall

Total summer precipitation was above average in both 1997 and 1998 (Table 2) (the weather station was not installed during 1996). Rain events that generated little or no runoff were far more frequent than larger, stream-fl ow-generating events, and provided the major-ity of precipitation (Table 3). Comparison with long-term inten-sity–duration–frequency curves for the Southwest (U.S. Weather Bureau, 1955) shows that there were no exceptionally intense storms during our study period. Rainfall came as two types of events: on fringes of convection thunderstorms or as frontal systems (Tuan et al., 1973). Reid et al. (1999) noted a similar dichot-omy of rainfall events in a study at Los Alamos National Laboratory, about 250 km northeast of Zuni. Long-term daily precipitation data from Zuni (Balling and Wells, 1990) shows that nearly 90% of annual precipitation comes as minor,

non-runoff-generating events (Fig. 4). Syed et al. (2003), working at the USDA Walnut Gulch Experimental Watershed near Tombstone, AZ, found that storms yielding low amounts of rainfall were much more frequent than high-yielding events. They also emphasized the importance of the storm core (defi ned as the portion of the storm having intensities >25 mm h−1) in generating runoff, which is rela-tively small in areal extent and therefore less likely to occur on smaller watersheds (Syed et al., 2003). Once a storm core reaches a watershed, however, the smaller the watershed the more likely it will generate runoff (Goodrich et al., 1997).

Slope PositionBoth yield and composition of sediment varied signifi cantly

by slope position (Tables 4 and 5), but yield and composition of runoff water did not (Norton, 2000). Included in the analy-ses are 104 samples from 28 summer storms at the 20-m2 slope position plots. Runoff coeffi cients (millimeters runoff/millime-ters precipitation × 100) averaged 3.8% for all the plot-years combined and ranged from 1.4 to 8.1% This compares favor-ably to nine hydrologic studies in pinyon–juniper environments reviewed by Wilcox (1994), in which runoff ranged from 1 to 23% of total precipitation, with all but one being 8% or less. Runoff coeffi cients for large storms in the U.S. Southwest are generally about 25% (Kirkby, 1978). Our values concur with our rainfall intensity data to show that there were no exception-ally intense storms during the study period.

Backslopes yielded more than twice as much sediment as SU and FS (Table 4). Backslopes also yielded the most C, N, and P in sediments, but concentrations of these soil organic matter (SOM)

Table 2. Approximate1997 and 1998 summer precipitation at the Weekoty study site.

Month Long-term average 19971997 percentage

of average1998†

1998 percentage of average

——— mm ——— % mm %May 12 31 263 1.7 14

June 8.9 52 582 0.2 2

July 53 39 75 148 281August 60 67 113 19 32Total 133 189 142 169 127

† 1998 data collected only through 19 August.

Fig. 4. Distribution of summer (April–September) daily precipitation, 1896 to 1985, Zuni Pueb-lo (as compiled by Balling and Wells, 1990).

Table 3. Total rainfall and yield and mean C/N ratio for sediment and organic material components in four peak rainfall–intensity categories†.

Peak intensity n‡

Total rainfall

Sediment Total N Total C Total PAvailable

PMean C/N

mm h−1 mm g m−2 ————— mg m−2——————0–20 193 158 20.2 36.9 533 6.1 0.27 14.6 a§21–40 38 113 388.5 431.3 7495 63.7 3.05 16.2 ab41–60 9 68 89.0 119.0 2166 22.0 0.77 17.6 bc>60 6 22 223.5 217.6 3813 43.2 1.21 18.1 c

† Sums and mean C/N ratio from all six runoff plots in Weekoty watershed, 1997 and 1998 summer seasons.

‡ n = total number of rainfall events recorded at the tipping bucket rain gauge.

§ Numbers within each column are signifi cantly different (P = 0.05) from those followed by different letters (LSD test).

480 SSSAJ: Volume 71: Number 2 • March –April 2007

constituents in BS sediments were equivalent to or less than SU and FS sediments (Table 5). This may be a result of relatively erosive fl ows due to steeper slopes, fi ner textured soils, and lack of ground cover (see Fig. 3), which remove proportionally more mineral soil materials from BS, while relatively gentle fl ows on the other two positions, particularly summits, move proportionally more low-density, relatively undecomposed surface litter organic material (OM) fragments.

These data support the work of Norton et al. (2003): stable soil microbial environments on SU positions, disturbance-driven erosional environments on BS, and depositional environments on FS. The relatively high erosion rates on BS may contribute to min-eralizing soil microbial environments with higher concentrations of mineral C, N, and P in SOM and also to deposition of mineralizing OM on FS, although we did not measure the movement of such materials among the slope positions. Conversely, summit positions are relatively stable with respect to erosion, which probably contrib-utes to relatively stable organic matter dynamics, as described by Norton et al. (2003), where effi cient nutrient cycling in an immo-bilizing soil microbial environment leads to low concentrations of mineral C, N, and P in SOM.

Cover TypeCover type impacted both yield and composition of sediment

washed from the 1-m2 plots (Tables 6 and 7), but not that of runoff water. Included in the analyses are 154 sediment and runoff samples collected from the 1-m2 soil cover plots during the summers of 1997 and 1998. Both bare soil and microbiotic crust cover yielded signifi cantly more sediment than pinyon and grass plots (Table 6). Oak and juniper plots yielded intermediate amounts of sediment. Sediment texture followed much the same trend as yield, with the

highest yielding cover types generally having the coarsest sediments. Except for oak plots, yields of C, N, and P in sediments depended largely on overall sediment yield. Sediments from oak cover were very rich in SOM, as indicated by very high yields of C and N. Based on sediment yield, microbiotic crust sediments had lower C/N and C/P ratios than those from other cover types, possibly due to decomposing algae and lichen from crusts in the sediments.

Concentrations (Table 7) of both organic C and total N in sed-iments from the oak plots were signifi cantly higher than other cover types. Microbiotic soil crust sediments had signifi cantly lower C, N, and P concentrations than all the cover types except bare soil.

Yields of sediment and runoff were considerably higher from the 1-m2 cover plots than from the 20-m2 slope position plots for the same rain events. This discrepancy may be due to a scale effect (Wilcox et al., 2003) in which longer slope lengths in the larger plots create transmission losses during the frequent small rainfall events.

Rainfall–Sediment InteractionsMost rain at the Weekoty watershed in the 1997 and 1998

fi eld seasons fell in low-intensity events (<20 mm h−1) that did not generate runoff or produce sediment (Table 3). Moderate-inten-sity rains of 21 to 40 mm h−1, with apparently the most erosive combination of frequency and power, produced by far the most sediment and OM components (C, N, and P). Storms with higher than 60 mm h−1 peak intensity were rare but produced much more sediment per event (37 g m−2 average) than the mid- (10 g m−2 for both 21–40 and 41–60 mm h−1 ranges) and low-intensity storms (0.10 g m−2). Sediment C/N ratios increased steadily with peak intensity, which suggests that lower intensity events transport higher quality, more decomposed OM.

Table 4. Mean summer runoff and sediment yield by slope position, 1996 to 1998, based on averages of sums from each plot for each fi eld season†.

Slope position‡ n§ Runoff Sediment Sand Silt Clay Total C Total N Total P Avail. P C/N¶ C/P¶ Avg. P/TP

L m−2 g m−2 –––––%–––––– ————— mg m−2—————

SU 8 1.9 a# 44 a 66 b 21 a 13 a 890 a 49.6 a 12.4 a 0.44 a 15 a 95 b 0.04 b

BS 8 2.8 a 125 b 45 a 30 b 25 b 2150 b 116.2 b 34.2 b 0.77 a 15 a 53 a 0.02 a

FS 6 2.3 a 57 a 58 b 27 ab 16 ab 1200 ab 73.9 ab 18.7 a 0.55 a 12 a 64 a 0.03 a

† Yields represent minimum yields from each slope position; at least one runoff event per year was missed at each trap.

‡ SU = summit; BS = backslope; FS = footslope.

§ n = total plot-years monitored in each slope position (see Table 1).

¶ C/N and C/P values represent means of individual C/N and C/P in samples where both C and N concentrations or C and P concentra-tions were available.

# Numbers within each column are signifi cantly different (P = 0.05) from those followed by different letters (LSD test).

Table 5. Concentration of nutrients in sediments by slope position†.

Slope position n‡ Total C Total N Total P Available P C/N§ C/P§ Avg. P/TP§

—————g kg−1————— mg kg−1

SU 34 31.4 b¶ 1.88 b 0.328 a 11.6 a 17.5 ab 119.0 b 0.0538 b

BS 39 23.4 a 1.37 a 0.336 a 9.8 a 17.6 b 74.4 a 0.0288 aFS 31 23.3 a 1.51 ab 0.310 a 10.4 a 15.9 a 79.3 a 0.0365 a

† SU = summit; BS = backslope; FS = footslope. Means of all samples within each slope position from the three watershed study areas.

‡ n = total number of samples collected in each slope position.

§ C/N and C/P values represent means of individual C/N and C/P in samples where both C and N concentrations or C and P concentrations were available. TP = total phosphorus.

¶ Numbers within each column are signifi cantly different (P = 0.05) from those followed by different letters (LSD test).

SSSAJ: Volume 71: Number 2• March –April 2007 481

Sediment yield at the Weekoty watershed 20-m2 plots is weakly correlated with runoff yield, but strongly correlated with rainfall intensity (P < 0.005) for storms we measured in 1997 and 1998 (Norton, 2000). This suggests that, with respect to sediment (but not runoff) yield, rainfall intensity is more infl uential than depth, duration, frequency, time to maximum, and the other rainfall parameters we calculated. Therefore total sediment yield at the 20-m2 plots is a reasonable indicator of runoff erosive power, total rain-fall notwithstanding. Relationships between sediment composition and sediment yield indicate how combined storm intensity affects the composition of materials moving from hillslopes. Figures 5a and 5b show negative logarithmic relationships between sediment yield and concentrations of organic C and total N, suggesting again that frequent, low-intensity rains move sediments with higher OM con-tents, while infrequent larger events move more sediment. Figure 5c shows a positive logarithmic relationship between sediment yield and C/N ratio, suggesting that more highly decomposed material is transported by lower intensity rainfall events. These data show that OM concentration and composition in low-intensity events are highly variable (low correlation coeffi cients), while those of higher intensity events are consistently low. Our results suggest that many low-intensity events yield sediments with high concentrations of high-quality OM, whereas higher intensity events always yield sedi-ments with low concentrations of low-quality OM. This emphasizes the importance of low-intensity rainfall events in nutrient cycling and transport (Sala and Lauenroth, 1982).

Considered within the context of Zuni’s annual precipitation pattern, these data may indicate pulses of OM decomposition and movement at the beginning of the runoff season, usually in early

July, which is a critical time for moisture and nutrient delivery to crops and natural vegetation. Zuni winters have freezing nighttime temperatures, warm daytime temperatures, and low-intensity rain or snow from frontal systems that generate little runoff (Tuan et al., 1973; Western Regional Climate Center, 2006). This combina-tion creates frequent freeze–thaw cycles that accelerate both physi-cal and microbial breakdown of forest litter (Honeycutt, 1995). In early spring, warmer daytime temperatures intensify freeze–thaw and bring drying–rewetting cycles that also enhance decomposition (Cui and Caldwell, 1997). Hot, dry conditions in May and June desiccate surface soils and organic debris. This winter–spring–early summer pattern may physically break down OM accumulated on slopes since the previous rainy season, leading to a fl ush of OM with relatively low C/N at the beginning of summer rainy seasons (Fig. 6). Wilcox (1994) reported a similar seasonal effect: erodibility of soils in pinyon–juniper interspaces was highest at the end of the winter freeze–thaw cycles and lowest at the end of the summer rainy season. As the rainy season progressed, slopes became more and more depleted of decomposed (low C/N) material.

Integration of the areal extent of slope positions and cover from Norton et al. (2003) with runoff data from this study (Table 8) shows estimated total sediment production from each slope posi-tion map unit in the three watersheds. These values show that exten-sion of transect and runoff plot data to whole watersheds ampli-fi es the effects of BS in all except the Laate site, where BS make up a smaller portion of the watershed. In the Sanchez watershed, the SU position is large because of bare exposed bedrock on the mesa top that was classifi ed as shoulder.

Table 7. Concentration of nutrients in sediments by cover type†.

Cover n‡ Organic C Total N Total P Available P C/N§ C/P§ Avg. P/TP§

—————g kg−1——————— mg kg−1

Bare soil 27 24.4 ab¶ 1.35 ab 0.281 ab 12.4 ab 17 b 87 ab 44.2 abMicrobiotic crust 30 12.9 a 0.91 a 0.246 a 7.1 a 14 ab 52 a 28.8 aOak 26 82.0 c 3.98 c 0.355 b 25.2 c 20 c 231 c 71.0 abJuniper 25 54.4 d 2.77 d 0.272 ab 25.8 c 19 cd 200 bd 95.1 abPinyon 22 43.8 bd 2.46 d 0.320 ab 20.1 bc 18 bd 137 cd 62.7 abGrass 24 38.5 bd 2.17 bd 0.334 ab 17.7 bc 18 bd 115 bd 52.8 b† Means of all samples collected from each vegetation cover type.

‡ n = total number of samples collected in each cover type.

§ C/N and C/P values represent means of individual C/N and C/P in samples where both C and N concentrations or C and P concen-trations were available. TP = total phosphorus.

¶Numbers within each column are signifi cantly different (P = 0.05) from those followed by different letters (LSD test).

Table 6. Mean summer runoff and sediment yield and composition by vegetation cover type. Yields are based on averages of sums from each plot for each fi eld season†.

Cover n‡ Runoff Sediment Sand Silt Clay Total C Total N Total P (TP) Avail. P C/N C/P Avg. P/TP

L m−2 g m−2 ————%———— ————— mg m−2——————Bare soil 4 9.4 a§ 486 b 82 a 12 a 6 a 7870 ab 412 ab 62.7 b 2.33 a 18 b 92 ac 0.04 a

Microbiotic crust

4 8.1 a 441 bc 78 ab 15 ab 7 a 5028 a 338 ab 54.4 bc 1.33 a 14 a 51 a 0.03 a

Oak 4 12.4 a 402 ab 63 c 26 c 12 b 32 600 b 1527 b 45.5 ab 7.69 a 20 b 217 b 0.07 a

Juniper 4 8.3 a 168 ab 76 abc 15 ab 9 a 4192 a 209 a 17.0 ac 1.74 a 20 b 178 bc 0.04 a

Pinyon 4 8.9 a 153 ac 61 cd 22 abc 17 c 3699 a 194 a 9.5 a 1.32 a 17 ab 123 ab 0.05 aGrass 4 9.4 a 107 a 67 bc 23 bc 10 ab 3296 a 181 a 15.0 ac 1.23 a 17 ab 108 ac 0.07 a

† Yields represent minimum yields from each slope position; at least one runoff event per year was missed at each trap.

‡ n = 4 based on two plots per cover type × 2 yr.

§ Numbers within each column are signifi cantly different (P = 0.05) from those followed by different letters (LSD test).

482 SSSAJ: Volume 71: Number 2 • March –April 2007

CONCLUSIONSOur data suggest that the unique combination of lithology,

topography, and rainfall pattern underlie transport and transforma-tion of sediments and OM that, in turn, may affect soil productiv-ity on alluvial fans of the Zuni area. Infrequent, intense storms are responsible for shaping the landscape, and frequent minor events may play pivotal roles in processing forest fl oor organic matter, sus-taining plant growth, and creating accumulation of materials on lower slopes that may set the stage for major movement of materials during larger events.

The results of this research contribute to the understand-ing of watershed hillslope characteristics that underlie the sustainability of runoff agriculture practiced by Zuni farmers for many centuries. Relationships among soil, landform, and vegetation patterns point to the wooded BS as a driving force behind movement and processing of runoff, sediment, and

OM. Backslopes’ large areal extent, steepness, and slowly per-meable subsoils covered by pinyon–juniper–oak forests yield OM-rich sediments that may be responsible for sustained pro-ductivity of alluvium-derived soils lower in watersheds.

Rainfall patterns dominated by minor events that generate short-distance sheet fl ow but not stream fl ow, create an OM-pro-cessing system that stimulates mineralization as it mixes materials and carries them downslope. Less-frequent fl ushing fl ows move accumulated sediments and OM from lower slopes to alluvial land-forms where traditional agricultural fi elds are typically located.

Traditional alluvial fan farming remains an important cul-tural activity among the Zuni and other Native American tribes in the Southwest, but the importance of hydrological connectivity between upland hillslopes and alluvial landforms extends beyond its value to traditional agriculture. Sediments, OM, and runoff from hillslopes are valuable resources for sustaining ecological diversity and productivity in fl oodplains, riparian areas, alluvial fans, and downstream aquatic systems. Functional fl oodplains and alluvial fans store sediments, attenuate peak fl ows, and absorb run-off that can maintain downstream perennial fl ows. Without func-tional hydrological connectivity between hillslopes and alluvial landforms (due to channel incision or channelization), products of hillslope erosion are lost—to become environmental liabilities in a feedback spiral of soil degradation where sediments and con-stricted fl ows damage downstream aquatic systems (Bull, 1997). Channel incision, or arroyo cutting, is a major issue across the Southwest (Cooke and Reeves, 1976; Elliot et al., 1999). This research improves our understanding of the movement and com-position of materials eroding from hillslopes, which, in intact allu-vial systems, sustain crucial cultural, ecological, and hydrological functions of alluvial fans and fl oodplains.

ACKNOWLEDGMENTSThis work was funded by National Science Foundation Grant no. DEB-9528458. A Research Assistantships for Minority High School Students supplemental award funded Kate Brown and her Zuni high school science class to monitor sediment traps. We are indebted to the Zuni Sustainable Agriculture Project, the Zuni Conservation Project, and the Zuni Tribe.

Fig. 5. Relationship between sediment yield and concentrations of C and N in sediments. *,** Signifi cant at 0.05 and 0.01 levels, respectively.

Fig. 6. Sediment C/N for the 1996 to 1998 rainy season (July–September) from 20-m2 runoff plots plotted against Day of the Year of the storm. Yields of total sediment, organic material, and particle-size distribution do not follow signifi -cant trends with respect to date.

SSSAJ: Volume 71: Number 2• March –April 2007 483

We thank Jeff Homburg, Todd Carlson, Marnie Criley, and Clara Wheeler for laboratory analysis and Troy Lucio, Lindsay Quam, Roman Pawluk, and participants in Zuni’s Job Training Partnerships Act (JTPA) program for fi eld assistance. We are grateful to Tom DeLuca, Urszula Norton, and Stephen Siebert for technical advice and editorial review, and to the research team, including Deborah Muenchrath, Stephen Williams, Pete Stahl, and Mark Ankeny.

REFERENCESBalling, R.C.J., and S.G. Wells. 1990. Historical rainfall patterns and arroyo

activity within the Zuni River drainage basin, New Mexico. Ann. Am. Assoc. Geogr. 80:603–617.

Bull, W.B. 1997. Discontinuous ephemeral streams. Geomorphology 19:227–276.

Cooke, R.U., and R.W. Reeves. 1976. Arroyos and environmental change in the American Southwest. Oxford Univ. Press, New York.

Cui, M., and M.M. Caldwell. 1997. A large ephemeral release of nitrogen upon wetting of dry soil and corresponding root responses in the fi eld. Plant Soil 191:291–299.

Cushing, F.H. 1920. Zuni breadstuff. Museum of the American Indian, Heye Foundation, New York.

Damp, J.E., S.A. Hall, and S. Smith. 2002. Early irrigation on the Colorado Plateau near Zuni Pueblo, New Mexico. Am. Antiq. 67:665–676.

Dick, W.A., and M.A. Tabatabai. 1977. An alkaline oxidation method for determination of total phosphorus in soils. Soil Sci. Soc. Am. J. 41:511–514.

Elliot, J.G., A.C. Gellis, and S.B. Aby. 1999. Evolution of arroyos: Incised channels of the southwestern United States. p. 153–185. In S.E. Darby and A. Simon (ed.) Incised river channels. John Wiley & Sons, New York.

Ferreira, V.A. 1990. Temporal characteristics of aridland rainfall events. p. 584–589. In R.H. French (ed.) Hydraulics/Hydrology of Arid Lands: Proc. Int. Symp. Am. Soc. Civ. Eng., Fort Collins, CO.

Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis. p. 383–411. In A. Klute (ed.) Methods of soil analysis. Part 1. Physical and mineralogical methods. 2nd ed. SSSA Book Ser. 5. SSSA, Madison, WI.

Gellis, A.C. 1998. Characterization and evaluation of channel and hillslope erosion on the Zuni Indian Reservation, New Mexico, 1992–95. Water Resour. Invest. Rep. 97–4281. U.S. Geol. Surv., Albuquerque, NM.

Goodrich, D.C., L.J. Lane, R.M. Shillito, S.N. Miller, K.H. Syed, and D.A. Woolhiser. 1997. Linearity of basin response as a function of scale in a semiarid watershed. Water Resour. Res. 33:2951–2965.

Homburg, J.A., J.A. Sandor, and J.B. Norton. 2005. Anthropogenic infl uences on Zuni agricultural soils. Geoarchaeology 20:661–693.

Honeycutt, W.C. 1995. Soil freeze-thaw processes: Implications for nutrient cycling. J. Minn. Acad. Sci. 59:9–14.

Kirkby, M.J. 1978. Hillslope hydrology. John Wiley & Sons, New York.Leopold, L.B., W.W. Emmett, and R.M. Myrick. 1966. Channel and hillslope

processes in a semiarid area, New Mexico. Prof. Pap. 352-G. U.S. Geol. Surv., Washington, DC.

Muenchrath, D.A., M. Kuratomi, J.A. Sandor, and J.A. Homburg. 2002. Observational study of maize production systems of Zuni farmers in semiarid New Mexico. J. Ethnobiol. 22:1–33.

Norton, J.B. 2000. Agroecology, hydrology, and conservation of ephemeral streams and alluvial fans, Zuni Pueblo, New Mexico. Publ. no. AAT 9993970. Univ. of Montana, Missoula.

Norton, J.B., F.J. Bowannie, P. Peynetsa, W. Quandelacy, and S.F. Siebert. 2002. Native American methods for conservation and restoration of semiarid ephemeral streams. J. Soil Water Conserv. 57:250–258.

Norton, J.B., J.A. Sandor, and C.S. White. 2003. Hillslope soils and organic matter dynamics within a Native American agroecosystem on the Colorado Plateau. Soil Sci. Soc. Am. J. 67:225–234.

Olsen, S.R., and L.E. Sommers. 1982. Phosphorus. p. 403–430. In A.L. Page et al. (ed.) Methods of soil analysis. Part 2. Chemical and biological methods. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.

Pawluk, R.R. 1995. Indigenous knowledge of soils and agriculture at Zuni Pueblo,

New Mexico. M.S. thesis. Iowa State Univ., Ames.Peterson, B.J., W.M. Wollheim, P.J. Mulholland, J.R. Webster, J.L. Meyer, J.L. Tank,

E. Marti, W.B. Bowden, H.M. Valett, A.E. Hershey, W.H. McDowell, W.K. Dodds, S.K. Hamilton, S. Gregory, and D.D. Morrall. 2001. Control of nitrogen export from watersheds by headwater streams. Science 292:86–90.

Reid, K.D., B.P. Wilcox, D.D. Breshears, and L. MacDonald. 1999. Runoff and erosion in a pinion–juniper woodland: Infl uence of vegetation patches. Soil Sci. Soc. Am. J. 63:1869–1879.

Sala, O.E., and W.K. Lauenroth. 1982. Small rainfall events: An ecological role in semiarid regions. Oecologia 53:301–304.

Sandor, J.A., J.B. Norton, J.A. Homburg, D.A. Muenchrath, C.S. White, S.E. Williams, C.L. Havener, and P.D. Stahl. 2007. Biogeochemical studies of a Native American runoff agroecosystem. Geoarchaeology (in press).

Schumm, S.A., and M.P. Mosley. 1973. Slope morphology. Dowden, Hutchinson, & Ross, Stroudsburg, PA.

Selby, M.J. 1993. Hillslope materials and processes. 2nd ed. Oxford Univ. Press, Oxford, UK.

Syed, K.H., D.C. Goodrich, D.E. Myers, and S. Sorooshian. 2003. Spatial characteristics of thunderstorm rainfall fi elds and their relation to runoff. J. Hydrol. 271:1–21.

Tuan, Y., C.E. Everard, J.G. Widdison, and I. Bennett. 1973. The climate of New Mexico. State Planning Offi ce, Santa Fe, NM.

U.S. Weather Bureau. 1955. Rainfall intensity–duration–frequency curves. Tech. Pap. no. 25. Hydrol. Serv. Div., U.S. Weather Bur., Washington, DC.

Western Regional Climate Center. 2006. Zuni, New Mexico—Climate summary. Available at www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?nmzuni (verifi ed 22 Nov. 2006). Western Regional Climate Center, Desert Res. Inst., Reno, NV.

Wilcox, B.P. 1994. Runoff and erosion in intercanopy zones of pinyon–juniper woodlands. J. Range Manage. 47:285–295.

Wilcox, B.P., D.D. Breshears, and H.J. Turin. 2003. Hydraulic conductivity in a pinon–juniper woodland: Infl uence of vegetation. Soil Sci. Soc. Am. J. 67:1243–1249.

Williams, J.D., and J.C. Buckhouse. 1991. Surface runoff plot design for use in watershed research. J. Range Manage. 44:411–412.

Table 8. Estimated total summer runoff and loss of sediment and nutri-ents by slope position for the three study sites on total area basis†.

Landform Area Runoff Sediment Total C Total N Total P Avail. P

ha m3 Mg —————kg————— g

All watersheds SU 71 1350 106 1658 83 18 632 BS 75 2245 442 7508 354 58 1843 FS 36 814 68 757 44 15 304Laate SU 2 37 4 33 2 1 13 BS 1 29 6 30 2 1 10 FS 2 43 4 39 2 1 15Sanchez SU 23 446 46 621 31 8 227 BS 18 514 107 1229 60 17 323 FS 17 388 39 407 23 8 153Weekoty SU 46 867 57 1004 50 10 392 BS 56 1703 333 6250 293 41 1510 FS 17 383 24 311 18 6 136

† Values for slope position were corrected for vegetation by: (i) dividing each vegetation cover plot value by the grass plot values (30% cover equal to slope position cover); (ii) multiplying slope position values by each of these cover-type conversion factors; (iii) multiplying these slope posi-tion–vegetation estimates by total area of each vegetation type within each slope position; and (iv) multiplying these corrected rates by the total area of each slope position in each watershed (from Norton et al., 2003).

‡ SU = summit; BS = backslope; FS = footslope.