Ecohydrology & Hydrobiology 16 (2016) 125–137

Contents lists available at ScienceDirect

Ecohydrology & Hydrobiology

journal homepage: www.elsevier.com/locate/ecohyd

Original Research Article

Model analysis of check dam impacts on long-term sediment

and water budgets in Southeast Arizona, USA
Laura M. Norman a,*, Rewati Niraula b
a
U.S. Geological Survey, Western Geographic Science Center, 520 N. Park Avenue, Suite #102K, Tucson, AZ 85719-5035, United States
b
University of Arizona, Hydrology and Water Resources, Tucson, AZ 85719, United States

A R T I C L E I N F O A B S T R A C T

Article history: The objective of this study was to evaluate the effect of check dam infrastructure on soil
Received 18 July 2015 and water conservation at the catchment scale using the Soil and Water Assessment Tool
Accepted 15 December 2015 (SWAT). This paired watershed study includes a watershed treated with over 2000 check
Available online 24 December 2015
dams and a Control watershed which has none, in the West Turkey Creek watershed,
Southeast Arizona, USA. SWAT was calibrated for streamflow using discharge documented
Keywords:
during the summer of 2013 at the Control site. Model results depict the necessity to
Soil and Water Assessment Tool (SWAT)
eliminate lateral flow from SWAT models of aridland environments, the urgency to
Calibration
Conservation standardize geospatial soils data, and the care for which modelers must document altering
Check dams parameters when presenting findings. Performance was assessed using the percent bias
Aridlands (PBIAS), with values of 2.34%. The calibrated model was then used to examine the impacts
of check dams at the Treated watershed. Approximately 630 tons of sediment is estimated to
be stored behind check dams in the Treated watershed over the 3-year simulation, increasing
water quality for fish habitat. A minimum precipitation event of 15 mm was necessary to
instigate the detachment of soil, sediments, or rock from the study area, which occurred 2% of
the time. The resulting watershed model is useful as a predictive framework and decision-
support tool to consider long-term impacts of restoration and potential for future restoration.
ß 2015 European Regional Centre for Ecohydrology of the Polish Academy of
Sciences. Published by Elsevier Sp. z o.o. All rights reserved.

1. Introduction oxygen (Branson et al., 1981; Gray et al., 2000). Sediment

can also serve as a transport mechanism for other
Soil erosion by water is one of the most important land pollutants. Rockström et al. (2010) argue that investments
degradation processes in aridland environments and is in catchment-scale management of water and soils is
linked to flooding and water-resource management necessary to minimize the risk for climate-related failures
(Poesen and Hooke, 1997). Sediment transport in streams and emphasize water harvesting systems to build resil-
is a major nonpoint source pollutant (NPS) in surface ience and address trade-offs between ecosystem services.
waters that compromises water clarity and health by Studies of rock structures placed in channels document
increasing water temperatures and decreasing dissolved decreased peak flows (Baker et al., 1995; Peterson et al.,
1960; Norman et al., 2010a, 2010b, 2015), a reduction in
sediment load (Hadley and McQueen, 1961; Hassanli et al.,
* Corresponding author. Tel.: +1 520 670 5510; fax: +1 520 670 5113.
2008; Polyakov et al., 2014), and increased vegetation
E-mail addresses: lnorman@usgs.gov (L.M. Norman), (Bombino et al., 2009; DeBano and Heede, 1987; DeBano
rewatin@gmail.com (R. Niraula). and Schmidt, 1990; Heede, 1978; Heede and DeBano,

http://dx.doi.org/10.1016/j.ecohyd.2015.12.001
1642-3593/ß 2015 European Regional Centre for Ecohydrology of the Polish Academy of Sciences. Published by Elsevier Sp. z o.o. All rights reserved.
126 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

1984; Hendrickson and Minckley, 1985; Norman et al., Negative effects of accelerated erosion and sedimenta-
2014). tion on water quality have been well-documented in the
Kingsford (2000) suggests better models of ecological Madrean Sky Islands (Lopes and Ffolliott, 1992; Marsh,
and hydrological impacts of dam installation for an 1968) and in other environments (Gray et al., 2000; Marsh,
improved understanding of the interaction between river 1968; Zabaleta et al., 2007). Soil and water conservation
flows, floodplain ecology, and management practices. experts have developed Best Management Practices
Traditional models cannot consider all the watershed- (BMPs) to reduce soil loss and improve water quality
scale erosion and sediment transport processes at once due (Young et al., 2010). Check dams and rock gabions are
to knowledge and/or data limitations (De Vente and examples of BMPs that can help to stabilize channels and
Poesen, 2005). However, there are many models that have trap sediment in upstream deposits. These deposits
been applied to develop estimates of runoff, erosion, and provide a nutrient-rich foundation that helps instigate
sediment yield and to mimic how check dams and rock plant establishment and promotes growth (DeBano and
structures might influence those (Martı́n-Rosales et al., Schmidt, 1990). By trapping sediment, these features also
2007; Remaı̂tre et al., 2008; Boix-Fayos et al., 2008; help improve water quality downstream, through attenu-
Norman et al., 2010a, 2010b; Garbrecht et al., 2014). In this ation of particles and pollutants (Pettersson, 1998).
study, we simulate the hydrologic processes documented Griffiths and Walton (1978) suggest 80–100 milligrams
in riparian areas treated with check dams using the Soil of solids per liter (mg/L) is the upper tolerance level for
and Water Assessment Tool (SWAT; Arnold et al., 1998; fish.
Neitsch et al., 2009) to extend the hydrologic budget and It is difficult to discern if changes in fluvial-sediment
describe the fate and transport of sediments. In selecting a load is the result of changes in management (BMPs,
model to use, our goals were to extrapolate on our limited grazing, etc.), climate, or if they are part of natural cycles
field data, be sparing with the adjustment of parameters, triggered when a geomorphic threshold is exceeded
provide better estimates of the response, and contribute to (Osterkamp, 1999). The objective of this research is to
the scientific understanding of the impacts of check dams provide a predictive framework to analyze impacts of
in aridlands. restoration in a 3-year model iteration which could be
In valleys of these arid and semiarid rangelands, fluvial projected into the future and to guide the further
sediment deposits accumulate on low parts of hillslopes restoration of ecological processes. The installation of
or in the channel and floodplains (Osterkamp, 1999; check dams is assessed in the context of demonstrating the
Coblentz, 2005). Gullies are formed by flowing water benefit of these features for protecting water quality. We
eroding soil on a hillside and leads to the destruction of hypothesize that the installation of a series of check dams
riparian habitat in headwater channels (DeBano and can support storage of organic soils and carbon, improve
Schmidt, 1990). Material is easily carried along when water quality, and increase local water supplies. Our
runoff begins on a hillslope but this declines after the first primary results include annual estimates of runoff,
flush, as new sediment must be detached (Zabaleta et al., evapotranspiration, soil–water storage, soil erosion, and
2007). Rates of sediment delivery in aridlands are hard to watershed sediment yield at a paired watershed site.
document due to the ephemeral nature of local streams
and large flood-recurrence intervals (Griffiths et al., 2006). 2. Materials and methods
The U.S. Department of Agriculture – Agricultural
Research Service (USDA-ARS) Southwest Watershed 2.1. Study area
Research Center (SWRC) is monitoring water and soil at
two heavily-instrumented locations in the Madrean Sky In the mid-western slope of the Chiricahua Mountains,
Islands: the Walnut Gulch Experimental Watershed in Southeast Arizona, average annual precipitation is
(WGEW) and the Santa Rita Experimental Range (SRER). 534 mm, with approximately 70% occurring July through
WGEW is a 149 km2 research area, approximately 312 mm September (Fuller, 2015). A historic stream gage (1919–
of precipitation is measured annually and elevation 1925) documented average annual daily flow at West
ranges 730 m (1220–1950 m; Goodrich et al., 2008). Turkey Creek 0.24 cubic meters per second (cms), which
Nearing et al. (2007) found that few precipitation events nearly doubled during the monsoon (US Geological Survey,
produce sediment yields, ranging 0.07–5.7 t/ha/yr, on 2015). This a valuable perennial stream and the largest
watersheds <5 ha and Nichols (2006) reported 0.17–1.1 t/ source of water contributing to both the Willcox Playa and
ha/yr on larger (35–159 ha) watersheds. Schreiber and groundwater Basin (Fig. 1), but lowering water levels
Kincaid (1967) documented runoff to be most dependent (Brown and Schumann, 1969; Jacobson et al., 2008) and
on storm size vs. antecedent moisture. The SRER in a total dissolved solids impact drinking water (Arizona
210 km2 study site where 250–500 mm of precipitation is Department of Water Resources, 2009).
measured annually and elevation ranges 500 m (900– Proprietors at El Coronado Ranch, who lease the West
1400 m; Polyakov et al., 2010). Lane et al. (1997) found Turkey Creek allotment from the U.S. Forest Service, began
rainfall extent and intensity, vegetative, soil cover, and installing low-lying rock check dams in 1983 averaging
topography influence sediment yield. Polyakov et al. >2.5 check dams per ha. Now, more than 2000 loose-rock
(2010) document approximately 6.4% of rainfall events structures have been strategically constructed by hand and
produce runoff, with the primary driver being the placed in tributaries flowing into the Turkey Pen sub-
intensity – sediment yield occurs in 26% of runoff events watershed (769 ha). The Turkey Pen is approximately 5 km
(0.85–6.69 t/ha/yr). long, with 554-m change in elevation. Sands, silts, and
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 127

Fig. 1. Map portraying the location of the West Turkey Creek hydrology in relationship to the Willcox, Playa, City of Willcox, Willcox Playa Watershed, and
the State of Arizona.

clays with organic material have deposited and filled areas 2.2. Model presentation and calibration
above the dams, forming wide areas of alluvial deposits.
Norman et al. (2015) paired the adjacent and untreated 2.2.1. Model characteristics
(Control) Rock Creek watershed (2405 ha) to document The Soil and Water Assessment Tool (SWAT), developed
variations in discharge using a modified-Continuous Slope by USDA-ARS and Texas A&M University, is a physically-
Area method. The Rock Creek flows through deep channels, based, continuous time, watershed-modeling program
with large boulders, and over exposed bedrock and is that can predict yields of water, sediment, nutrients, and
approximately 10 km from outlet to peak, where topo- agricultural chemicals in large watersheds (Arnold et al.,
graphic relief varies by 1238 m. Results demonstrated a 1998; Neitsch et al., 2009). It is a public domain
lower runoff response (peak flow) in the Turkey Pen hydrological-transport model that can operate on a
(Treated) watershed, yet 28% more volume via extended daily-time step in basin scale. SWAT calculates surface
base-flows. runoff and accommodates for some transmission losses.
Photographs from the intersection of Turkey Pen Water that infiltrates the surface is divided into multiple
(Treated) to West Turkey Creek (Fig. 2a and b), and of layers for routing, and some moves through the soil via
the gages situated at Rock Creek (Control; Fig. 2c) and ‘lateral flow’ or unsaturated flow. The combination of
Turkey Pen (Treated; Fig. 2d) during high flows, demon- surface runoff, lateral flow, and return flow contributes to
strate reduced turbidity in treated channels, but water- the streamflow being estimated. A detailed description of
quality has not been documented. SWAT can be found in Neitsch et al. (2009). Calibration is
128 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

Fig. 2. Photographs display clarity of the Turkey Pen (Treated) vs. turbidity at the confluence with West Turkey Creek (untreated) (a) looking downstream
and Turkey Pen (Treated) on right (8/30/13; E. Gwilliam); (b) looking upstream where Turkey Pen (Treated) coming in on the left side (8/28/14; L. Norman).
Photographs of gages during high flow at (c) Rock Creek (Control; 8/14/14) and (d) Turkey Pen (Treated; 8/5/14; by F. Brinkerhoff).

suggested to improve model performance and increase the soil layer (%)]. All factors are determined by SWAT from the
reliability of flow predictions. Niraula et al. (2012a, 2015) input data or during the each model run (Q and q), except
found that it is important to calibrate the model spatially to for P, which defaults to 1. SWAT then estimates sediment
analyze the effect of land-use change but not as important using both lateral and groundwater sources (Chaubey
for climate change impacts. et al., 2006). Sediment yield is simulated for each unique
SWAT calculates erosion based on the Universal Soil Loss hydrological response unit, independent of relational
Equation (USLE; Eq. (1)), which estimates average annual landscape position within each subwatershed (White,
soil loss at the plot scale (Wischmeier and Smith, 1978): 2001; White et al., 2010).
SWAT is based on mechanisms derived from plot-scale
USLE : A ¼ RKLSCP (1) studies which present challenges for mimicking BMPs and
where A = predicted soil loss (tons/acre/year), R = rainfall scaling processes up to the watershed level (Arnold et al.,
and runoff factor, K = soil erodibility factor, LS = slope factor 1998; Gassman et al., 2007). SWAT has various surface-
(length and steepness), C = crop and cover management water impoundment tools (potholes, wetlands, ponds, and
factor, and P = conservation/support practice factor. The reservoirs) to simulate the impacts on hydrology and
amount of sediment to reach the channel can be calculated sediment and/or nutrients (Du et al., 2005) and these have
using a sediment delivery ratio (SDR) multiplied by been modified by many scientists in different applications
estimates of gross erosion (Williams, 1977). Many (Arnold et al., 2001; Almendinger et al., 2014; Kiesel et al.,
modifications of the USLE, some with geospatially derived 2010; Liu et al., 2008; Mishra et al., 2003, 2007; Ouessar
SDRs have been used to quantify sediment loads in SE et al., 2008; Tesfahunegn et al., 2013; Waidler et al., 2011;
Arizona (Brady et al., 2001; Norman, 2005, 2007; Norman Wang et al., 2008, 2010; Wang and Yu, 2012). Yang et al.
et al., 2007). SWAT substitutes a SDR by using a derived (2009) used SWAT to assess the flow diversion terrace
runoff factor instead of the rainfall factor (Bonumá et al., impacts on surface water quality in the Black Brook
2014) and applies the modified-USLE (MUSLE) to simulate Watershed, New Brunswick, Canada. Results indicated
estimates of sediment yield (Williams, 1975; Eq. (2)). flow diversion terraces contributed to the reduction of
sediment yield by 56%, and reduced water yield during the
Y ¼ 11:8ðQ surf qpeak areaHRU Þ0:56 KLSCPCRFG (2) summer growing seasons by 20%. In the Yanhe watershed
(7725 km2), Loess Plateau, China, 6572 check dams (1
where Y is the sediment yield (metric tons/day), Qsurf is the check dam/118 ha) have been installed beginning in 1970s.
surface runoff volume (mm H2O/ha), qpeak is the peak Xu et al. (2013) split data into rainy/dry seasons to account
runoff rate (m3/s), areaHRU is the area of the HRU (ha), and for monsoons in SWAT, finding that in rainy seasons (May–
CFRG is the coarse fragment factor [CFRG = exp October) runoff decreased by 16–30% and sediment
(0.053 * rock), where rock is the percent rock in the first decreased 35–86%, and in dry season (November–April),
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 129

runoff increased by 20–101%. Mwangi et al. (2015) used calibrated the outlet at Rock Creek (Control) and trans-
SWAT to evaluate multiple conservation practices in the ferred the calibrated parameters to the Turkey Pen, which
Sasumua Watershed, Kenya (107 km2), including filter allowed for the model to simulate flow there, as if there
strips, contour farming, parallel terraces, grassed water- were no check dams. We were then able to compare the
ways, and their combinations finding various reductions in differences between the observed hydrological and simu-
surface runoff and sediment yields along with slight lated data at Turkey Pen (Treated) on an annual basis and
increases in base flow. document the influence of check dams on the hydrologic
and sediment budgets.
2.2.2. Data used in this study for model calibration Soils, topographic, land-use, and daily weather data
Discharge data collected by Norman et al. (2015) from were collected from difference sources and assembled in a
June to October 2013, from the paired watershed hydro- geodatabase. Geospatial data were clipped to the study
graphs were used to calibrate the SWAT model. An effort area and converted to a common Universal Transverse
was made to document the location of all the check dams Mercator (UTM) projection. Soils data were mapped by the
in the main channels the Turkey Pen (Treated) watershed USDA Natural Resource Conservation Service (NRCS),
using a handheld Garmin global positioning system (GPS) collected at scales ranging from 1:12,000 to 1:250,000,
receiver (Fig. 3), but after GPSing almost 2000 check dams, representative of the dominant component in the mapping
we still observe many more tucked into side drainages and unit. However, the higher resolution Soil Survey Geo-
along every hillslope. We realized that the density of check graphic Database (SSURGO) dataset has poor representa-
dams has altered the channel characteristics too drastically tion at our study area, because NRCS does not map USFS
to portray as surface-water impoundments. Therefore, we land at the same scale as private lands. Oddly, the older and

Fig. 3. Map portraying hydrology with digitized (GPSed) check dams classified by elevation, in the Turkey Pen Watershed, where the top map inset depicts
the location in West Turkey Creek.
130 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

generally lower-resolution State Soil Geographic SWAT discretized channels and flow paths of the
(STATSGO) dataset had better spatial resolution in our system using a minimum threshold area (25 ha) to define
study area. Furthermore, polygons describing soil have a stream. Eleven points were added to represent tributaries
little to no tabular data and geospatial properties from the collected using GPS in the field. Two outlets were
STATSGO soil map and the soil database downloaded with documented at the confluences of both the paired water-
ArcSWAT 2009 were manually merged (Winchell et al., sheds (Turkey Pen and Rock Creek) where the tributaries
2009). The highest-resolution digital elevation models were monitored in the summer of 2013, and 81 sub-basins
(DEM) were found at 1/3 arc-second (10 m) and down- (averaging 0.39 km2) were defined. We used a threshold
loaded from the National Elevation Dataset (NED). The value of 10% for land use, soil and slope in defining
USGS National Land Cover Database (NLCD2006) is a 16- hydrologic response units (HRUs), ultimately generating
class land-cover classification based on the Landsat 666 total HRUs in ArcSWAT Version 2012.10_1.6 (released
Enhanced Thematic Mapper+ (ETM+) from 2006 at a 2/11/13; Winchell et al., 2009) and was run for 3 years (8/
spatial resolution of 30 m (Wickham et al., 2014). Daily 26/11–9/30/14).
precipitation and temperature data for two weather
stations located in or at the perimeters of, the combined 3. Results and discussion
32 km2 study area were assimilated for input to the
model. The West Turkey Creek ALERT gage 3040 Seasonal discharge for the monsoon season in 2013,
(318510 36.0000 N, 1098200 9.0000 W) is approximately 0.6 km developed by Norman et al. (2015), served as the basis of
south of the edge of the Turkey Pen (Treated) at 1907 m the calibration. First the model was calibrated to accom-
elevation and the Long Park ALERT gage 3090 modate the water balance and then the shape of the
(318530 46.3000 N, 1098170 0.3000 W) is at the peak of the Rock hydrograph. At the Control site, only 2½% of precipitation
Creek (Control), elevation 2768 m (Fuller, 2015). These is observed to be allocated to surface runoff in the channel
event-based tipping buckets report in real-time whenever (Norman et al., 2015). Almost all flow in aridlands is
there is 1 mm of precipitation. surface runoff (vs. subsurface or lateral flow) and we
There is no observed sediment data for which to adopted a customized script for the modeling in the Santa
calibrate the sediment predictions in the model. However, Cruz Watershed (developed by Niraula et al., 2015), to
we know that the magnitude, duration, and frequency of account for the very-low flow in the unsaturated zone due
the flows will dictate the behavior of bedload or suspended to rainfall. A rigorous manual calibration approach was
sediment load (Gray et al., 2000; Marsh, 1968; Zabaleta adopted to further refine the SWAT model for the study
et al., 2007). Therefore, we relied solely on the calibration area based on a review of current literature that identified
of the flow regime. We pushed all sediment parameters parameters commonly adjusted (Arnold et al., 2001). Prior
and routing variables to their maximum options to test the knowledge from past studies in the region was also relied
sensitivity of these to predicted sediment deposition to upon for possible parameter ranges and the most sensitive
create a qualitative measure, but sediment yield from the parameters (Niraula et al., 2012a, 2015). After the model
HRU that is transported into the main channel during the results mimicked the observed water budget, the calibra-
time step (SYLD) predictions resulted in negligible changes tion was focused to match the rainfall–runoff response to
in every output file. We attribute this to the fact that SYLD storms, time lags, and travel times and imitate the shape of
is predicted using generated runoff in SWAT (vs. precipi- the observed data’s hydrograph in terms of peaks,
tation) and our adjustments to the code minimize lateral recession, and antecedent moisture conditions. An infor-
flows. mal sensitivity analysis was carried out by varying
A sediment delivery ratio (SDR) was applied based on parameters one at a time to create a qualitative measure
drainage area and multiplied by the soil (gross) erosion of parameter sensitivities (Niraula et al., 2012b). The final
predicted using the USLE for determining average annual calibration consisted of adjusting only the coefficients
sediment yield. Vanoni (1975) used the data from listed in Table 1.
300 watersheds throughout the world to develop a Hyeto-hydrographs portray average precipitation (mm)
generalized equation to estimate SDR (Eq. (3)). in relationship to both observed and the modeled
discharge (cms) at Rock Creek (Control; Fig. 4), and Turkey
SDR ¼ 0:4724A0:125 ; where A Pen (Treated; Fig. 5).
2 In order to examine the average tendency of the
¼ watershed area ðkm Þ (3)
simulated data to deviate from the observed, we applied a
percent bias (PBIAS) calculation (Eq. (5); Gupta et al.,
The USDA Soil Conservation Service (USDA-SCS, 1975)
1999):
developed an equation to calculate SDR based on the data
from the Blackland Prairie, Texas (Eq. (4)):
ðaverage of observedaverage of simulatedÞ100
PBIAS ¼
average of observed
SDR ¼ 0:51A0:11 ; where A
(5)
¼ drainage area in square miles: (4)
where error in average flows at Rock Creek is calculated as
While these are not standard SWAT equations, we 2.34% and in Turkey Pen is 119.76%. Low-magnitude
applied them outside of the model to estimate a watershed values at Rock Creek indicate accurate model simulation in
SDR and resulting sediment yield. the model for the study area if there are no dams. The large
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 131

Table 1
SWAT model parameters included in the calibration and final adjustment.

Parameter Description File Default Range Calibrated value

ESCO Soil evaporation compensation factor .bsn 0.95 0–1 0.1

TRNSRCH Fraction of transmission losses from main channel that enter deep aquifer .bsn 0 0–1 0.36
GW_REVAP Groundwater revap coefficient .gw 0.02 0.02–0.2 0.2
REVAPMN Threshold depth of water in shallow aquifer .gw 1 0–500 0
GW_DELAY Ground water delay time (days) .gw 31 0–500 15
CN2 Initial Soil Conservation Service runcurve number for moisture condition II .mgt Varies 35–82 0.75
CH_K(2) Effective hydraulic conductivity in main channel alluvium (mm/h) .rte 0 0.01–500 10
SOL_AWC Available water capacity of the soil .sol 0–1 0–1 1.5
USLE_K USLE equation soil erodibility (K) factor .sol Varies 0–0.65 0.25
USLE_C USLE C factor for water erosion applicable to the land cover/plant plant.dat 0.003 0.001–0.5 0.25

negative value at Turkey Pen demonstrates huge overestima- 2013 and compared their associated runoff with the
tion bias because the treatment (check dams) is not included in model’s estimates (Fig. 6). In the first storm event (and at
the model, thus demonstrating the effect they have on the beginning of the hyeto-hydrograph; Fig. 5), the model
hydrology. Other metrics to report include R2 and Nash– is over-estimating runoff in Turkey Pen (treated) with little
Sutcliffe Efficiency (NSE), but these were developed to examine difference in Rock Creek (the control). This is because the
error on a monthly time step (Moriasi et al., 2007), and would initial demands are greater in Turkey Pen when first flows
not reflect the accuracy of our calibrated model in mimicking are captured. Over time (during the next two precipitation
the daily peaks/volumes documented on the hydrograph. events), the model starts to under-estimate runoff at
Volume might decrease at the beginning of the monsoon in the Turkey Pen in comparison with Rock Creek. The model was
Treated watershed but will increase over time, based on the not calibrated to take into account the extended steady
observed dataset and associated analysis using PBIAS. flow from storage in the check dams or the increased
To better analyze how the hydrographs become volume realized over time in the Treated watershed.
desynchronized through time, we separated them into The model was run using the available measured
the three main precipitation events (storms) captured in precipitation data starting 8/26/11, for 3 years to simulate

Fig. 4. Hyeto-hydrograph portraying average precipitation (mm) with observed vs. model-simulated discharge (cms) at the Rock Creek (Control) site.

Fig. 5. Hyeto-hydrograph portraying average precipitation (mm) with observed vs. model-simulated discharge (cms) at the Turkey Pen (Treated) site.
132 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

Fig. 6. Graph of three largest precipitation events captured to compare the difference in output from observed to modeled runoff in the Treated vs. Control
watersheds.

annual water and sediment budgets. Results demonstrate out of 1146 days, only 20 days have precipitation that
comparatively low surface runoff, total water yield, results in eroded soil. The precipitation threshold needed
surface-runoff/total flow ratio (0.08), and groundwater to generate soil loss is 15 mm/day.
ratio (0.28). Average annual basin stress days include The soil loss per HRU was summed and averaged per
water stress days (63) and temperature stress days (50) subbasin to create estimated sediment budgets for each
due to the aridlands climate. We assume that the subwatershed (Fig. 8), where an average of 792 tons/year
differences identified in Fig. 6 radically alter the Treated of soil loss is predicted at Turkey Pen during the model
watershed’s budget, where runoff volume increases (28%) iteration vs. 3774 tons/year at Rock Creek.
over time via lateral and baseflow, supported by soil–water Using the equation developed by Vanoni (1975; Eq. (3)),
storage, the increased residence time of water on the a simple estimate for a SDR was calculated for the Turkey
landscape also increases recharge to the aquifer down- Pen watershed, where SDR = 0.61. Using the equation
stream. developed by the USDA-SCS (1975), another estimate of
Our efforts to mimic sediment yield using the MUSLE SDR was calculated, where SDR = 0.45. We multiplied the
did not provide results, but fortunately, SWAT also average soil loss by each estimated SDR, to make a crude
calculates the USLE equation which is completely depen- estimate of sediment yield. Approximately 356–483 tons
dent on precipitation (mm). The average soil loss of sediment would likely be yielded from the Turkey Pen
calculated using the USLE per HRU 8.9 tons/ha (Treated) watershed, given no management considerations
(Max = 147.9; Min = 0.02) during the time step calculated (i.e. if there were no check dams) and approximately 1428–
(8/26/11–9/30/14). We summed the estimates of soil loss 1936 tons are being yielded at the Control (Rock Creek).
and water yield to better examine cycles through time Hadley and McQueen (1961) showed that structures in
(Fig. 7). All soil loss occurs in 2% (1.75%) of the timeframe; Wyoming reduced sediment loads by up to 75% and

Fig. 7. Graph portraying precipitation events that result in model-predicted soil erosion in the WTC paired watershed, with associated water yields.
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 133

Fig. 8. Map depicting the estimated average soil loss (tons/year) per subbasin per subwatershed, predicted over the 3-year simulation.

Boix-Fayos et al. (2008) calculated 77% retention at check 4. Conclusions

dams in the Rogativa catchment, Murcia, Spain. Yang et al.
(2009) and Xu et al. (2013) saw similar results using the This research explored the impacts of land-use man-
SWAT model. Yang et al. (2009) found a reduction of water agement and restoration using check dams on fluvial-
yield (summer) of 20% and sediment yield 56%, in the sediment load and water quality. Field data collected over
Black Brook Watershed, New Brunswick, Canada. Xu et al. a three-month period was used to calibrate and train a
(2013) found check dams reduced runoff in rainy season watershed model to simulate 3 years’ worth of hydrologic
16–30%, and then increased in the dry season 20–101%, and sediment cycles. Obstacles we encountered were both
and reduced sediment yield 35–86% in the Yanhe intrinsic to the model and unavailability of input. Modeling
watershed, China. Based on these studies and using values is much faster and less expensive than implementing long-
from Polyakov et al. (2014), who documented that check term data collection but allows for more error. Our
dams would retain 50% of the sediment yield over a 3-year hydrologic calibration is based on only one monsoon
study, we estimate the total soil retained in the Turkey Pen season and used to develop sediment transport estimates.
(Treated) stored by check dams throughout the watershed Higher-resolution input data (DEM and soils input) would
approximately 50% = 178–242 tons/year (Table 2). decrease error, along with more field measurements for
calibration and validation of the model, including water
quality (suspended sediment) and the expansion of flow.
Soils data from the NRCS are not developed at high
Table 2
resolutions for land owned and managed by the USFS.
Model results (assuming untreated): average discharge, soil loss, and
sediment yield at the study site watershed and estimates of treatment on However, the USFS does create their own very-detailed
sediment yield. Terrestrial Ecosystem Survey that ultimately could be
translated into NRCS-style geospatial datasets and greatly
Turkey Pen Rock Creek
(Treated) (Control)
benefit national watershed modeling. The lateral flow
component of the SWAT model that defaults to portray
Size (ha) 769 2405
sub-surface flow is not appropriate in aridlands and
Avg. flow (cms) 0.077 0.192
Average soil loss (USLE; tons/year) 792 3174 needed to be altered to better reflect dry antecedent
Avg. sediment yield 356–483 1428–1936 moisture conditions. This created errors in the model’s
(SDR = 0.45–0.61; tons/year) ability to calculate sediment yield, due to the dependency
Avg. sediment yield IF check dams 178–242 714–968
of the model on lateral flow that simulates runoff and
(ton/year)
instigates erosion. While we have no measurements of
134 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

sediment yield for comparison, anecdotal and pictorial over-withdrawals in the watershed and the inability to
evidence portrays sedimentation in the watershed and balance recharge currently – maybe by increasing resi-
data provided within the model outputs gave estimates of dence time and baseflow using check dams in the entire
soil loss – useful to approximate sediment yield and West Turkey Creek watershed, management could help
sediment retention. increase recharge (water quantity) and reduce impaired
The calibrated model simulated discharge with only a water quality. There is a need to better understand the
2% error for the Control watershed. More than 2000 check long-term potential and future impacts of sediment
dams installed in the Treated watershed created a huge storage related to fire, drought and biodiversity. Biodiver-
discrepancy in model outputs given the calibration to the sity is the measure of the variety of organisms present and
untreated (Control) site. Those variations depict the can be impacted by changes in ecosystem variation. The
influence that check dams have in the real world. There resulting watershed model demonstrates a predictive
is a measured reduction in peak flow and a decline in framework to analyze long-term impacts of restoration
volume at the beginning of the monsoon compared to the and also to display potential impacts of future restoration.
model estimates in the Treated watershed. This is followed For example, the Control watershed (Rock Creek) could be
by a relative increase in volume and duration of flow fitted with check dams improve water quality and quantity
subsequent to the monsoon. The desynchronized hydro- in the future. Subbasins predicted to generate higher soil
graphs produced by the model portray this effect. At the loss are mapped and could be targeted as hot spots for
beginning of the monsoon, the model over-estimates erosion control in future BMPs.
runoff at Turkey Pen (Treated) due to the initial demands of
created by check dams. Toward the end of the summer Conflict of interest
monsoon, the model starts to under-estimate flows at
Turkey Pen, as it was not calibrated to take into account the None declared.
soil–water storage in the check dams nor the increased and
extended baseflow realized over time. Ethical statement
The simulation predicted erosion and soil loss on
20 days of the 3-year simulation, dependent on minimum Authors state that the research was conducted accord-
precipitation of 15 mm/day. Based on the soil loss ing to ethical standards.
estimates, sediment yield from the study area is estimated
to be 356–483 tons/year from the Turkey Pen (Treated)
Acknowledgements
watershed (if there were no check dams) and 1428–
1936 tons/year at the Rock Creek (Control). Check dams
We dedicate this paper to our friend Joan Ruvinsky,
could retain 178–242 tons/year in the Turkey Pen
who, with her partner Kathleen Knipp, volunteered many
(Treated) watershed over the 3-year simulation. A detailed
hours on the trail collecting GPS points at hundreds of
survey of the amounts, locations and characteristics of the
check dams in the Turkey Pen Watershed. We also thank
soils behind dams, stratigraphy of the layers and even dating
Josiah and Valer Austin (Cuenca Los Ojos Foundation),
of materials would be useful for better understanding of
Brian Iserman and Peter Acton (JE Fuller Hydrology &
soil–water storage but could be a difficult given the nature
Geomorphology, Inc.), Valerie Gordon, and Gerry, Garrett,
and timing of the installation. New research to document
and Alana Norman for their help in collecting data. We
the forest and terrestrial carbon storage associated with
appreciate the careful peer reviews of this manuscript
restoration using gabions/check dams is also warranted. The
provided by James Callegary and Waite Osterkamp (USGS).
restoration being done is creating organic carbon storage in
Lastly, we want to recognize the valuable assistance for
the soils, currently undocumented, but useful for restoration
SWAT modeling provided by D. Phillip Guertin (University
managers and those interested in carbon storage (ecosystem
of Arizona), James Almendinger (St. Croix Watershed
service value increases). Model simulations could ultimately
Research Station, Science Museum of Minnesota), and
be used to evaluate sediment and organic material
Michael White (USDA-ARS). References to commercial
retention, which is very important for aquatic ecosystem
vendors of software products or services are provided
management and carbon sequestration, but rarely docu-
solely for the convenience of users when obtaining or using
mented in the literature.
USGS software. Such references do not imply any
The velocity and volume of water flow are important
endorsement by the U.S. Government.
factors impacted by the installation of check dams that
alter total suspended solids (TSS) or sediments in the Funding body
water. High concentrations of TSS cause water quality This research was conducted with support from the
impairment and create problems for aquatic life. Based on Land Change Science (LCS) Program, under the Climate and
the discharge and sediment yield estimates developed Land Use Change (CLU) Mission Area of the U.S. Geological
herein, we estimate that Turkey Pen (Treated) averages Survey (USGS).
86 mg/L vs. the Rock Creek (Control) site is estimated to
have 278 mg/L. There is a small population of Yaqui chub References
(Gila purpurea), a native fish that inhabit clear-water pools
of small streams, in the Treated (Turkey Pen) watershed. Almendinger, J.E., Murphy, M.S., Ulrich, J.S., 2014. Use of the soil and water
assessment tool to scale sediment delivery from field to watershed in
In considering the dependency of the Willcox basin on an agricultural landscape with topographic depressions. J. Environ.
nearby mountain-front recharge, it is important to note the Qual. 43 (1), 9, http://dx.doi.org/10.2134/jeq2011.0340.
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 135

Arizona Department of Water Resources, 2009. Section 3.14 Willcox Basin Goodrich, D.C., Keefer, T.O., Unkrich, C.L., Nichols, M.H., Osborn, H.B.,
Arizona Water Atlas, vol. 3, pp. 534–602. Retrieved from http://www. Stone, J.J., Smith, J.R., 2008. Long-term precipitation database, Walnut
azwater.gov/AzDWR/StatewidePlanning/wateratlas/SEArizona/ Gulch Experimental Watershed, Arizona, United States. Water
documents/Volume_3_WIL_final.pdf. Resour. Res. 44 (5), W05S04, http://dx.doi.org/10.1029/
Arnold, J.G., Allen, P.M., Morgan, D.S., 2001. Hydrologic model for design 2006WR005782.
and constructed. Wetlands 21 (2), 167–178, http://dx.doi.org/ Gray, J.R., Glysson, G.D., Turcios, L.M., Schwarz, G.E., 2000. Comparability
10.1672/0277-5212(2001)021[0167:HMFDAC]2.0.CO;2. of suspended-sediment concentration and total suspended solids
Arnold, J.G., Srinivasan, R., Muttiah, R.S., Williams, J.R., 1998. Large area data (Water-Resources Investigations Report No. 00-4191) USGS.
hydrologic modeling and assessment. Part I: Model development. J. Griffiths, P.G., Hereford, R., Webb, R.H., 2006. Sediment yield and runoff
Am. Water Resour. Assoc. 34 (1), 73–89, http://dx.doi.org/10.1111/ frequency of small drainage basins in the Mojave Desert, U.S.A.
j.1752-1688.1998.tb05961.x. Geomorphology 74 (1–4), 232–244, http://dx.doi.org/10.1016/j.geo-
Baker Jr., M.B., DeBano, L.F., Ffolliott, P.F., 1995. Hydrology and watershed morph.2005.07.017.
management in the Madrean Archipelago. In: DeBano, L.H., Ffolliott, Griffiths, W.H., Walton, B., 1978. The effects of sedimentation on the
P.H., Ortega-Rubio, A., Gottfried, G.J., Hamre, R.H., Edminster, C.B., aquatic biota (Alberta Oil Sands Environmental Research Program No.
tech. coords., Biodiversity and Management of the Madrean Archi- 35) , pp. 86.
pelago: The Sky Islands of Southwestern United States and North- Gupta, H.V., Sorooshian, S., Yapo, P.O., 1999. Status of automatic calibra-
western Mexico, Tucson, AZ, September 19–23, 1994. U.S. tion for hydrologic models: comparison with multilevel expert cali-
Department of Agriculture, Forest Service, Rocky Mountain Forest bration. J. Hydrol. Eng. 4 (2), 135–143.
and Range Experiment Station, Fort Collins, CO, pp. 329–337. Hadley, R.F., McQueen, I.S., 1961. Hydrologic effects of water spreading in
Boix-Fayos, C., de Vente, J., Martı́nez-Mena, M., Barberá, G.G., Castillo, V., Box Creek basin, Wyoming (No. WSP-1532-A) United States Geologi-
2008. The impact of land use change and check-dams on catchment cal Survey. Retrieved from http://pubs.er.usgs.gov/publication/
sediment yield. Hydrol. Process. 22 (25), 4922–4935, http:// wsp1532A.
dx.doi.org/10.1002/hyp.7115. Hassanli, A.M., Nameghi, A.E., Beecham, S., 2008. Evaluation of the effect
Bombino, G., Gurnell, A.M., Tamburino, V., Zema, D.A., Zimbone, S.M., of porous check dam location on fine sediment retention (a case
2009. Adjustments in channel form, sediment calibre and vegetation study). Environ. Monit. Assess. 152 (1–4), 319–326, http://dx.doi.org/
around check-dams in the headwater reaches of mountain torrents, 10.1007/s10661-008-0318-2.
Calabria, Italy. Earth Surf. Process. Landf. 34 (7), 1011–1021, http:// Heede, B.H., 1978. Designing gully control systems for eroding water-
dx.doi.org/10.1002/esp.1791. sheds. Environ. Manag. 2 (6), 509–522, http://dx.doi.org/10.1007/
Bonumá, N.B., Rossi, C.G., Arnold, J.G., Reichert, J.M., Minella, J.P., Allen, BF01866709.
P.M., Volk, M., 2014. Simulating landscape sediment transport capac- Heede, B.H., DeBano, L.F., 1984. Gully rehabilitation – a three-stage
ity by using a modified SWAT model. J. Environ. Qual. 43 (1), 55, process in a sodic soil. Soil Sci. Soc. Am. J. 48 (6), 1416–1422.
http://dx.doi.org/10.2134/jeq2012.0217. Hendrickson, D.A., Minckley, W.L., 1985. Cienegas: vanishing climax
Brady, L.M., Gray, F., Wissler, C., Guertin, D.P., 2001. Spatial variability of communities of the American Southwest. Desert Plants (USA). Re-
sediment erosion processes using GIS analysis within watersheds in a trieved from http://agris.fao.org/agris-search/search.do?f=1986/US/
historically mined region, Patagonia Mountains, Arizona. U.S. Geo- US86099.xml;US8603277.
logical Survey Open-File Report, 01-267. Retrieved from http:// Jacobson, J., Davis, T., Hinckley, A., Schmerge, D., Flora, S., 2008. Water
geopubs.wr.usgs.gov/open-file/of01-267/ level changes in Willcox Basin, Arizona 1999–2005 (Water Level
Branson, F.A., Gifford, G.F., Renard, K.B., Hadley, R.F., 1981. Rangeland Change Map Series Report No. 1) Arizona Dept of Water Resources.
hydrology. Soc. Rng. Mgmt. Rng. Sci. Ser. 1 . Kingsford, R.t., 2000. Ecological impacts of dams, water diversions and
Brown, S.G., Schumann, H.H., 1969. Geohydrology and Water Utilization river management on floodplain wetlands in Australia. Austral Ecol.
in the Willcox Basin, Graham and Cochise Counties Arizona (U.S. 25 (2), 109–127, http://dx.doi.org/10.1046/j.1442-
Geological Survey Water-Supply Paper No. 1859-F) . Retrieved from 9993.2000.01036.x.
http://pubs.usgs.gov/wsp/1859f/report.pdf. Kiesel, J., Fohrer, N., Schmalz, B., White, M.J., 2010. Incorporating land-
Chaubey, I., Migliaccio, K.W., Green, C.H., Arnold, J.G., Srinivasan, R., 2006. scape depressions and tile drainages of a northern German lowland
Phosphorus modeling in soil and water assessment tool (SWAT) catchment into a semi-distributed model. Hydrol. Process. 24 (11),
model. In: Modeling Phosphorus in the EnvironmentCRC Press, Boca 1472–1486, http://dx.doi.org/10.1002/hyp.7607.
Raton, FL, pp. 163–188. Lane, L.J., Hernandez, M., Nichols, M., 1997. Processes controlling sedi-
Coblentz, D., 2005. The tectonic evolution of the Madrean Archipelago and ment yield from watersheds as functions of spatial scale. Environ.
its impact on the geoecology of the Sky islands. In: Gottfried, G.J., Model. Softw. 12 (4), 355–369, http://dx.doi.org/10.1016/S1364-
Gebow, B.S., Eskew, L.G., Edminster, C.B., compilers, Connecting 8152(97)00027-3.
Mountain Islands and Desert Seas: Biodiversity and Management Liu, Y., Yang, W., Wang, X., 2008. Development of a SWAT extension
of the Madrean Archipelago II, Tucson, AZ, May 11–15, 2004. U.S. module to simulate riparian wetland hydrologic processes at a wa-
Department of Agriculture, Forest Service, Rocky Mountain Research tershed scale. Hydrol. Process. 22 (16), 2901–2915, http://dx.doi.org/
Station, Fort Collins, CO, pp. 62–68. Retrieved from http://www.fs.fed. 10.1002/hyp.6874.
us/rm/pubs/rmrs_p036/rmrs_p036_062_068.pdf. Lopes, V.L., Ffolliott, P.F., 1992. Hydrology and watershed management of
De Vente, J., Poesen, J., 2005. Predicting soil erosion and sediment yield at oak woodlands in southwestern Arizona. In: Ecology and Manage-
the basin scale: scale issues and semi-quantitative models. Earth Sci. ment of Oak and Associated Woodlands: Perspectives in the South-
Rev. 71 (1–2), 95–125, http://dx.doi.org/10.1016/j.ear- western United States and Northern MexicoRocky Mountain Forest
scirev.2005.02.002. and Range Experiment Station, Fort Collins, CO, pp. 224. Retrieved
DeBano, L.F., Heede, B.H., 1987. Enhancement of riparian ecosystems with from http://www.alibris.com/Ecology-and-management-of-
channel structures. J. Am. Water Resour. Assoc. 23 (3), 463–470, oak-and-associated-woodlands-perspectives-in-the-southwestern-
http://dx.doi.org/10.1111/j.1752-1688.1987.tb00824.x. United-States-and-northern-Mexico-April-27-30-1992-Sierra-
DeBano, L.F., Schmidt, L.J., 1990. Potential for enhancing riparian habitats Vista-Arizona-Peter-F-Ffolliott/book/1877526.
in the southwestern United States with watershed practices. For. Ecol. Marsh, J.A., 1968. The effect of suspended sediment and discharge on
Manag. 33–34, 385–403, http://dx.doi.org/10.1016/0378- natural infiltration of ephemeral streams. (M.S.)University of Ari-
1127(90)90205-P. zona, Tucson, AZ. Retrieved from http://arizona.openrepository.com/
Du, B., Arnold, J.G., Saleh, A., Jaynes, D.B., 2005. Development and appli- arizona/bitstream/10150/191501/1/
cation of SWAT to landscapes with tiles and potholes. Trans. ASAE 48 azu_td_hy_e9791_1968_245_sip1_w.pdf.
(3), 1121–1133, http://dx.doi.org/10.13031/2013.18522. Martı́n-Rosales, W., Gisbert, J., Pulido-Bosch, A., Vallejos, A., Fernández-
Fuller, J., 2015. Cochise County ALERT system. Retrieved from http:// Cortés, A., 2007. Estimating groundwater recharge induced by engi-
jefullerdata.com/Cochise/index1h.html. neering systems in a semiarid area (southeastern Spain). Environ.
Garbrecht, J.D., Nearing, M.A., Shields, F.D., Tomer, M.D., Sadler, E.J., Bonta, Geol. 52 (5), 985–995, http://dx.doi.org/10.1007/s00254-006-0541-5.
J.V., Baffaut, C., 2014. Impact of weather and climate scenarios on Mishra, A., Fröbrich, J., Kar, S., 2003. Potentials and applicability of the
conservation assessment outcomes. J. Soil Water Conserv. 69 (5), SWAT model in check dam management in small watershed. Re-
374–392, http://dx.doi.org/10.2489/jswc.69.5.374. trieved from http://swat.tamu.edu/docs/swat/conferences/2003/
Gassman, P.W., Reyes, S.H., Green, C.H., Arnold, J.G., 2007. The Soil and Session%20D/Ashok.pdf.
Water Assessment Tool: historical development, applications, and Mishra, A., Froebrich, J., Gassman, P., 2007. Evaluation of the SWAT model
future research directions, vol. 50(4), pp. 1211–1250, Presented at the for assessing sediment control structures in a small watershed in
Trans. ASABE. India. Am. Soc. Agric. Biol. Eng. 50 (2), 469–477.
136 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137

Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., Poesen, J.W.A., Hooke, J.M., 1997. Erosion, flooding and channel manage-
Veith, T.L., 2007. Model evaluation guidelines for systematic quanti- ment in Mediterranean environments of southern Europe. Prog. Phys.
fication of accuracy in watershed simulations. Trans. ASABE 50 (3), Geogr. 21 (2), 157–199, http://dx.doi.org/10.1177/
885–900, http://dx.doi.org/10.13031/2013.23153. 030913339702100201.
Mwangi, J.K., Shisanya, C.A., Gathenya, J.M., Namirembe, S., Moriasi, D.N., Polyakov, V.O., Nearing, M.A., Nichols, M.H., Scott, R.L., Stone, J.J.,
2015. A modeling approach to evaluate the impact of conservation McClaran, M.P., 2010. Long-term runoff and sediment yields from
practices on water and sediment yield in Sasumua Watershed, Kenya. small semiarid watersheds in southern Arizona. Water Resour. Res. 46
J. Soil Water Conserv. 70 (2), 75–90, http://dx.doi.org/10.2489/ (9), W09512, http://dx.doi.org/10.1029/2009WR009001.
jswc.70.2.75. Polyakov, V.O., Nichols, M.H., McClaran, M.P., Nearing, M.A., 2014. Effect
Nearing, M.A., Nichols, M.H., Stone, J.J., Renard, K.G., Simanton, J.R., 2007. of check dams on runoff, sediment yield, and retention on small
Sediment yields from unit-source semiarid watersheds at Walnut semiarid watersheds. J. Soil Water Conserv. 69 (5), 414–421,
Gulch. Water Resour. Res. 43 (6), W06426, http://dx.doi.org/10.1029/ http://dx.doi.org/10.2489/jswc.69.5.414.
2006WR005692. Remaı̂tre, A., van Asch, T.W.J., Malet, J.-P., Maquaire, O., 2008. Influence of
Neitsch, S.L., Arnold, J.G., Kiniry, J.R., Williams, J.R., 2009. Soil and Water check dams on debris-flow run-out intensity. Nat. Hazards Earth Syst.
Assessment Tool: Theoretical Documentation. Retrieved from http:// Sci. 8 (6), 1403–1416, http://dx.doi.org/10.5194/nhess-8-1403-2008.
www. brc.tamus.edu/swat/ Rockström, J., Karlberg, L., Wani, S.P., Barron, J., Hatibu, N., Oweis, T.,
Nichols, M.H., 2006. Measured sediment yield rates from semiarid range- Bruggeman, A., Farahani, J., Qiang, Z., 2010. Managing water in rainfed
land watersheds. Rangel. Ecol. Manag. 59 (1), 55–62, http:// agriculture – the need for a paradigm shift. Agric. Water Manag. 97
dx.doi.org/10.2111/05-075R1.1. (4), 543–550, http://dx.doi.org/10.1016/j.agwat.2009.09.009.
Niraula, R., Meixner, T., Norman, L.M., 2015. Determining the importance Schreiber, H.A., Kincaid, D.R., 1967. Regression models for predicting on-
of model calibration for forecasting absolute/relative changes in site runoff from short-duration connective storms. Water Resour. Res.
streamflow from LULC and climate changes. J. Hydrol. 522, 439– 3 (2), 389–395, http://dx.doi.org/10.1029/WR003i002p00389.
451, http://dx.doi.org/10.1016/j.jhydrol.2015.01.007. Tesfahunegn, G.B., Vlek, P.L., Tamene, L., 2013. Application of SWAT model
Niraula, R., Norman, L.M., Meixner, T., Callegary, J., 2012a. Multi-gauge to assess erosion hotspot for sub-catchment management at Mai-
calibration for modeling the semi-arid Santa Cruz watershed in Negus catchment in northern Ethiopia. East Afr. J. Sci. Technol. 2 (2),
Arizona–Mexico border area using SWAT. Air Soil Water Res. (5), 97–123.
41, http://dx.doi.org/10.4137/ASWR.S9410. USDA Soil Conservation Service (USDA-SCS), 1975. Sediment Sources,
Niraula, R., Kalin, L., Wang, R., Srivastava, P., 2012b. Determining nutrient Yields, and Delivery Ratios. National Engineering Handbook, Section
and sediment critical source areas with SWAT model: effect of 3 – Sedimentation.
lumped calibration. Trans. ASABE 55 (1), 147–157. US Geological Survey, 2015. USGS Water-Data Site Information for the
Norman, L.M., 2005. Modeling land use change and associate water Nation. Retrieved from http://waterdata.usgs.gov/nwis/si.
quality impacts in the Ambos Nogales watershed, United States– Vanoni, V.A., 1975. Sedimentation engineering (Manual and Report No.
Mexico border. The University of Arizona. Retrieved from http:// 54) American Society of Civil Engineers, New York, NY. Retrieved from
phdtree.org/pdf/25545994-modeling-land-use-change-and- http://cedb.asce.org/cgi/wwwdisplay.cgi?139314.
associate-water-quality-impacts-in-the-ambos-nogales-watershed- Waidler, D., White, M., Steglich, E., Wang, S., Williams, J., Jones, C.A.,
united-states-mexico-border/ Srinivasan, R., 2011. Conservation practice modeling guide for SWAT
Norman, L.M., 2007. United States–Mexican border watershed assess- and APEX. Texas Water Resources Institute Technical Report (399) ,
ment: modeling nonpoint source pollution in Ambos Nogales. J. pp. 71.
Borderl. Stud. 22 (1), 79–97, http://dx.doi.org/10.1080/ Wang, L., Yu, J., 2012. Modelling detention basins measured from high-
08865655.2007.9695670. resolution light detection and ranging data: modelling detention
Norman, L.M., Brinkerhoff, F.C., Gwilliam, E., Guertin, D.P., Callegary, J.B., basins from high-resolution LiDAR data. Hydrol. Process. 26 (19),
Goodrich, D.C., Nagler, P.L., Gray, F., 2015. Hydrologic response of 2973–2984, http://dx.doi.org/10.1002/hyp.8314.
streams restored with check dams in the Chiricahua Mountains, Wang, X., Shang, S., Qu, Z., Liu, T., Melesse, A.M., Yang, W., 2010. Simulated
Arizona. River Res. Appl., http://dx.doi.org/10.1002/rra.2895. wetland conservation–restoration effects on water quantity and
Norman, L.M., Gray, F., Guertin, D.P., Wissler, C., Bliss, J.D., 2007. Tracking quality at watershed scale. J. Environ. Manag. 91 (7), 1511–1525,
acid mine-drainage in Southeast Arizona using GIS and sediment http://dx.doi.org/10.1016/j.jenvman.2010.02.023.
delivery models. Environ. Monit. Assess. 145 (1–3), 145–157, http:// Wang, X., Yang, W., Melesse, A.M., 2008. Using Hydrologic Equivalent
dx.doi.org/10.1007/s10661-007-0024-5. Wetland Concept within SWAT to Estimate Streamflow in Water-
Norman, L.M., Huth, H., Levick, L., Shea Burns, I., Phillip Guertin, D., Lara- sheds with Numerous Wetlands, vol. 51 (1). , pp. 55–72.
Valencia, F., Semmens, D., 2010a. Flood hazard awareness and hydro- White, M.J., 2001. Evaluation of Management Practices and Examination
logic modelling at Ambos Nogales, United States–Mexico border. J. of Spatial Detail Effects Using the SWAT Model. Oklahoma State
Flood Risk Manag. 3 (2), 151–165, http://dx.doi.org/10.1111/j.1753- University. Retrieved from https://afrsweb.usda.gov/sp2UserFiles/
318X.2010.01066.x. ad_hoc/62060505MikeWhite/pdfs/white_thesis.pdf.
Norman, L.M., Levick, L.R., Guertin, D.P., Callegary, J.B., Quintanar Gua- White, M.J., Storm, D.E., Busteed, P., Stoodley, S., Phillips, S.J., 2010.
darrama, J., Zulema Gil Anaya, C., Prichard, A., Gray, F., Octavio Evaluating conservation program success with Landsat and SWAT.
Gastelum Ceballos, F., 2010b. Nogales flood detention study. U.S. Environ. Manag. 45 (5), 1164–1174, http://dx.doi.org/10.1007/
Geological Survey Open-File Report, 2010-1262, pp. 112. s00267-010-9458-6.
Norman, L.M., Villarreal, M.L., Pulliam, H.R., Minckley, R., Gass, L., Tolle, C., Wickham, J., Homer, C., Vogelmann, J., McKerrow, A., Mueller, R., Herold,
Coe, M., 2014. Remote sensing analysis of riparian vegetation re- N., Coulston, J., 2014. The Multi-Resolution Land Characteristics
sponse to desert marsh restoration in the Mexican Highlands. Ecol. (MRLC) Consortium – 20 years of development and integration of
Eng. 70C, 241–254, http://dx.doi.org/10.1016/j.ecoleng.2014.05.012. USA national land cover data. Remote Sens. 6 (8), 7424–7441.
Osterkamp, W.R., 1999. Runoff and Sediment Yield from Proxy Records: Williams, J.R., 1975. Sediment-yield prediction with universal equation
Upper Animas Creek Basin, New Mexico (Rocky Mountain Research using runoff energy factor. In: Proceedings of the Sediment-Yield
Station Research Paper No. RMRS-RP-18) United States Department of Workshop, vol. Publ. ARS-S-40. USDA Sedimentation Laboratory,
Agriculture Forest Service, Fort Collins, CO, pp. 50. Oxford, MS, pp. 244–252.
Ouessar, M., Bruggeman, A., Mohtar, R., Ouerchefani, D., Abdelli, F., Williams, J.R., 1977. Sediment delivery ratios determined with sediment
Boufelgha, M., 2008. Future of drylands – an overview of evaluation and runoff models. Erosion and Solid Matter Transport in Inland
and impact assessment tools for water harvesting. In: Lee, C., Schaaf, Waters, IAHS-AISH Publication No. 122, pp. 168–179.
T. (Eds.), The Future of Drylands. Springer, Netherlands, pp. 255–267, Winchell, M., Srinivasan, R., Di Luzio, M., Arnold, J., 2009. ARCSWAT
Retrieved from http://link.springer.com/chapter/10.1007/ 2.3.4 Interface for SWAT2005: User’s Guide. Blackland Research
978-1-4020-6970-3_29. Center, Texas Agricultural Experiment Station, Temple, TX.
Peterson, H.V., DeJulio, O.P., Rupkey, R.H., 1960. Effect of Bureau of Land Wischmeier, W.H., Smith, D.D., 1978. Predicting Rainfall-Erosion Losses:
Management structures on discharge of San Simon Creek and on A Guide to Conservation Plannings (Agriculture Handbook No. 537)
irrigation diversions in Safford Valley, Arizona (U.S. Geological Survey USDA, Washington. Retrieved from http://topsoil.nserl.purdue.edu/
Administrative Report) , pp. 85. usle/AH_537.pdf.
Pettersson, T., 1998. Water quality improvement in a small stormwater Xu, Y.D., Fu, B.J., He, C.S., 2013. Assessing the hydrological effect of the
detention pond. Water Sci. Technol. 38 (10), 115–122, http:// check dams in the Loess Plateau, China, by model simulations. Hydrol.
dx.doi.org/10.1016/S0273-1223(98)00740-9.
 L.M. Norman, R. Niraula / Ecohydrology & Hydrobiology 16 (2016) 125–137 137

Earth Syst. Sci. 17 (6), 2185–2193, http://dx.doi.org/10.5194/hess-17- Young, K.D., Younos, T., Dymond, R.L., Kibler, D.F., Lee, D.H., 2010. Appli-
2185-2013. cation of the analytic hierarchy process for selecting and modeling
Yang, Q., Meng, F.-R., Zhao, Z., Chow, T.L., Benoy, G., Rees, H.W., Bourque, stormwater best management practices. J. Contemp. Water Res. Educ.
C.P.-A., 2009. Assessing the impacts of flow diversion terraces on 146 (1), 50–63, http://dx.doi.org/10.1111/j.1936-704X.2010.00391.x.
stream water and sediment yields at a watershed level using SWAT Zabaleta, A., Martı́nez, M., Uriarte, J.A., Antigüedad, I., 2007. Factors
model. Agric. Ecosyst. Environ. 132 (1–2), 23–31, http://dx.doi.org/ controlling suspended sediment yield during runoff events in small
10.1016/j.agee.2009.02.012. headwater catchments of the Basque Country. CATENA 71 (1), 179–
190, http://dx.doi.org/10.1016/j.catena.2006.06.007.