Special Section: Frontiers Spatiotemporal Patterns of ...plsullivan/thomasetal_2013.pdf · along with other chemical tracers to conceptualize catchment hydrology (Holmes, 2011; Jin

www.VadoseZoneJournal.org

Spatiotemporal Patterns of Water Stable Isotope Compositions at the Shale Hills Critical Zone Observatory: Linkages to Subsurface Hydrologic ProcessesTo better understand flow pathways and patterns in the subsurface, a stable isotope moni-toring network was established at the Susquehanna-Shale Hills Critical Zone Observatory (SSHCZO). Soil water samples were collected approximately biweekly using suction-cup lysimeters installed at multiple depths along four different transects in the catchment. Groundwater and stream water were collected daily in the valley using automatic samplers, while precipitation samples were collected automatically on an event basis. The 3+ years (2008–2012) of monitoring data showed strong seasonal precipitation isotope composi-tions, which were imprinted in seasonal patterns of soil water at different spatial locations and depths. The groundwater isotope composition remained relatively constant through-out the year and closely matched the yearly amount-weighted precipitation average, sug-gesting groundwater received recharge water in each season, although recharge mecha-nisms differed between growing and nongrowing seasons. Soil water samples showed clear attenuation with depth, with the largest variability in the shallow soil water (£30 cm) mir-roring precipitation inputs, moderate variability in the intermediate depths (40–100 cm), and the least variability in the deep soil water (³120 cm) where the average remained near the groundwater average. Soil water isotope composition profiles also provided clear evi-dence for preferential flow occurring both laterally and vertically in different seasons and at various soil depths in the catchment. Putting all together, the extensive dataset of soil water isotopic compositions obtained in this study have provided a number of insights into complex subsurface hydrologic processes that are transferable to other similar landscapes.

Abbreviations: LMWL, local meteoric water lines; SSHCZO, Susquehanna-Shale Hills Critical Zone Observa-tory; VSMOW, Vienna standard mean ocean water.

Knowledge of subsurface flowpaths and timescales is essential to the understanding of soil formation and hydrologic processes in the Critical Zone, the region between deep groundwater and the tree canopy. Soil water plays an essential role by trans-porting major constituents important for terrestrial ecosystem services including carbon, nitrogen, and other nutrients, chemical weathering products, sediments, and potential con-taminants. Furthermore, soil water links precipitation to groundwater, facilitating recharge and dictating when subsurface flow occurs. As anthropogenic forces continue to stress the environment, it becomes more important to understand this linkage and transport pro-cesses within the vadose zone. Stable isotopes of water provide an important environmental tracer for understanding vadose zone hydrology, groundwater recharge, mixing processes, stream flow generation, subsurface storage, and catchment yield (Vanclooster et al., 2005; Leibundgut et al., 2009).

Water stable isotopes (deuterium and 18O) allow for nondestructive, long-term monitor-ing of subsurface hydrology. Stable isotopes have been used to determine groundwater composition and recharge rates (Gat, 1971; Darling and Bath, 1988; Davisson and Criss, 1993; Criss and Davisson, 1996; Winograd et al., 1998), preferential flow pathways and the old water-new water paradox (McDonnell, 1990; Leaney et al., 1993; Kirchner, 2003; Vogel et al., 2010), soil water evaporation (Zimmermann et al., 1968; Barnes and Allison, 1983, 1988; Gazis and Feng, 2004), hydrograph separation (Sklash et al., 1976; Sklash and Farvolden, 1979; Rice and Hornberger, 1998), and sources of transpired water (Ehleringer and Dawson, 1992; Brooks et al., 2009; Gierke, 2012).

Spatiotemporal investigation of high-resolution δO and δD sampling at Susquehanna-Shale Hills Critical Zone Observatory (SSHCZO) revealed pat-terns of subsurface hydrology. Heavy damping of seasonal amplitude was found to occur with depth indicat-ing the importance of hydrodynamic mixing. Variations in seasonal soil isotope profiles suggest prevalence and mechanisms of preferential flow.E.M. Thomas, C.J. Duffy, P.L. Sullivan, and G.H. Holmes, Dep. of Civil and Environmen-tal Engineering, Pennsylvania State Univ., University Park, PA 16802; H. Lin, Dep. of Ecosystem Science and Management, Penn-sylvania State Univ., University Park, PA, 16802; S.L. Brantley, Earth and Environmen-tal Systems Inst., Pennsylvania State Univ., University Park, PA, 16802; L. Jin, Dep.of Geological Sciences, Univ. of Texas at El Paso, El Paso, TX, 79968. *Corresponding author ([email protected]).

Vadose Zone J. doi:10.2136/vzj2013.01.0029Received 20 Jan. 2013.

Special Section: Frontiers of Hydropedology in Vadose Zone Research

Evan M. Thomas* Henry Lin Christopher J. Duffy Pamela L. Sullivan George H. Holmes Susan L. Brantley Lixin Jin

© Soil Science Society of America 5585 Guilford Rd., Madison, WI 53711 USA.All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including pho-tocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

www.VadoseZoneJournal.org p. 2 of 16

The Susquehanna-Shale Hills Critical Zone Observatory (SSHCZO) was established in 2007 to determine regolith forma-tion and evolution as well as hydrologic flowpaths and timescales within a small, forested catchment. To date, many studies at the SSHCZO have begun to answer these questions by identifying the prominence of hydrologic processes including soil moisture dynamics (Lin, 2006; Lin et al., 2006; Lin and Zhou, 2008), solute transport (Jin et al., 2010, 2011; Andrews et al., 2011; Kuntz et al., 2011), and stream flow generation mechanisms (Lynch, 1976; Duffy, 1996). Additionally, water stable isotopes have been used along with other chemical tracers to conceptualize catchment hydrology (Holmes, 2011; Jin et al., 2011). Jin et al. (2011) used variations in Mg concentrations and water stable isotopes to iden-tify three types of water storage within a specific hillslope of the SSHCZO: (i) low-flow waters in soils where chemical dissolution occurs, (ii) high-flow waters in soils where residence times are short, and (iii) groundwater storage in fractured shale bedrock. Their findings indicate typically low-flow waters are stored within soil matrix, high-flow waters are found at the interfaces of low perme-ability, and groundwater is found continually in the valley floor. Additionally, Jin et al. (2011) used water isotopes to estimate that average residence times are less than a year, indicating the drainage of the hillslope is relatively fast.

Preferential flow pathways have been found to be prominent in the catchment (Lin, 2006; Lin and Zhou, 2008; Graham and Lin, 2011; Jin et al., 2011). Analysis of 175 precipitation events over 3 yr by Graham and Lin (2011) showed that preferential flow occurred during at least 90% of the events. While these studies have success-fully conceptualized the subsurface hydrology at the SSHCZO, water isotopes have not yet been used to investigate the temporal and spatial patterns of subsurface hydrologic processes across the catchment.

The goal of this study is to improve the understanding of the spa-tial and temporal drivers of subsurface hydrologic processes at the SSHCZO utilizing high frequency sampling of water stable iso-topes. Specifically, we examine the influence of hillslope orienta-tion, slope location, landform type, and soil depth on soil water isotope compositions to determine water flowpaths associated with specific soil or landscape features. We aim to improve the understanding of temporal patterns of water isotope compositions involving: seasonal differences (growing vs. nongrowing), snow-melt dynamics, and periods of water scarcity. This study will help improve the understanding of mechanisms involved in controlling spatiotemporal patterns of soil water isotope composition, evapo-transpiration patterns during growing seasons, and subsurface preferential flow dynamics.

6Materials and MethodsSite DescriptionLandform and VegetationThe SSHCZO, a 7.9-ha forested catchment located in central Pennsyl-vania, is characteristic of the Ridge and Valley Physiographic Province of the eastern United States. The V-shaped headwater catchment is oriented east–west and contains an intermittent first order stream that typically flows from late September to early June. The SSHCZO lies within the Shavers Creek watershed of the Juniata River, which in turn is a sub-basin of the Susquehanna River. Slopes within the catchment range from 25 to 48%, with elevation varying from 310 m at the ridge top to 256 m at the stream outlet (Lin, 2006). The strike and dip of the underlying shale bedrock are approximately N54°E and 76°NW, respectively, as measured in one outcrop exposed near the catchment valley (Jin et al., 2010) but shallower dip has been measured by televiewer in a borehole on near outlet of catchment the valley floor near outlet of catchment (Kuntz et al., 2011).

There are four distinct landform units within the catchment: north (south-facing) slope with deciduous forest and underbrush, south (north-facing) slope with deciduous forest and thicker underbrush, valley floor with evergreen (hemlock) trees along the western side and deciduous oak-hickory forest on the eastern side, and topo-graphic depressional areas (swales) with deciduous forest cover and deeper soils. There are seven swales distributed over the catchment, with five on the north slope and two on the south slope, each sepa-rated by near-planar slopes. The catchment was clear-cut approxi-mately 60 yr ago and the current mature forest is composed of deciduous trees on the slopes and ridges [Maple (Acer spp.), Oak (Quercus spp.), and Hickory (Carya spp.)] and eastern hemlock conifers on the valley floor [Tsuga Canadensis (L.) Carriere].

SoilsThe soil survey indicates that soils in this catchment were formed from shale residuum or colluvium and contain many channery shale fragments throughout most of the soil profiles (Lin et al., 2006). Hillslope soils have silt loam texture in general, moderately devel-oped soil structure, and are generally well drained. In the valley floor, there are redoximorphic features starting to appear at 30–60 cm in soil profiles. Throughout the catchment, there is about 0.05 m thick organic layer (Oe-horizon) (Lin et al., 2006). There are five distinct soil series that are distinguished based on depth to bedrock, land-form unit, and drainage condition (Fig. 1). The Weikert series is the most prevalent soil type (loamy-skeletal, mixed, active, mesic Lithic Dystrudepts) and is generally <0.5 m thick to bedrock. The Berks soil series (loamy-skeletal, mixed, active, mesic Typic Dystrudepts) is 0.5–1 m thick to bedrock. The Rushtown series (loam-skeletal over fragmental, mixed, mesic Typic Dystrochrepts) is located in the center of swales as well as in the eastern end of the catchment and is >1 m thick to bedrock. Soil series located on the valley floor are distinguished by a fragipan-like layer and the depth to redoxi-morphic features. The Ernest series (fine-loamy, mixed, superactive, mesic, Aquic Fragiudults) occupies the majority of the streambed


and contains redox features starting at 0.3–0.5 m. Th e Blairton series (fi ne-loamy, mixed, active, mesic Aquic Hapludults) is of minor extent and contains redox features at >0.6 m depth. Further information regarding the soil survey and features can be found in Lin et al. (2006). Specifi c soil characteristics, weathering reactions, soil chemistry, and mineralogy have been previously reported for the locations of the lysimeter nests described herein (Jin et al., 2010; Jin and Brantley, 2011; Ma et al., 2011).

ClimateTh e SSHCZO is located in a humid climate where the long-term average annual precipitation is 1006 mm, which is approximately evenly distributed throughout the year. Mean annual temperature is 10.1°C. Summers are typically hot and humid with temperatures >30°C and precipitation events consisting of short, high-intensity storms. Winters are typically wet and cold with temperatures less than −10°C and precipitation events consisting predominantly of snow (based on a 30-yr record, NOAA, 2007).

Data Collecti onsHydrologic MonitoringMeteorological data were collected in a small clearing area at the ridge top of the catchment from 2009 to 2012. Precipitation was

monitored using an Ott–Pluvio weighing bucket (Hach Company, Loveland, CO) in 10-min intervals with a precision of 0.1 mm and Laser Precipitation Monitor (LPM, Th ies Clima, Göttingen, Ger-many) in 10-min intervals with precision of 0.01 mm. Th e LPM is capable of determining the type of precipitation (e.g., drizzle, rain, snow, or hail) based on the size and velocity of the precipi-tation. Data were collected and transmitted on-line in real time and processed annually for missing data. Missing data were fi lled directly using 16 tipping-bucket rain gages with HOBO recorders (Davis Instruments Corp., Hayward, CA) located throughout the catchment and the LPM. Th e tipping-buckets record rainfall at a height of 0.3 m with a precision of 0.2 mm. Because the tipping buckets were typically located under the canopy and only collected throughfall, these data were used only when necessary.

A double V-notch weir located at the outlet of the stream was used to monitor stream discharge accurately during high and low fl ows. Water depths were recorded in 1-min intervals, integrated to 10 min values and converted to discharge using a rating curve devel-oped by Nutter (1964). Depth to groundwater was monitored at one well in the stream riparian zone from January 2009 to Decem-ber 2011 in 10-min intervals using Druck pressure transducers (Campbell Scientifi c Inc., Logan, UT) (Fig. 1).

Fig. 1. Topographic map of Susquehanna-Shale Hills Critical Zone Observatory (SSHCZO) with stable isotope monitoring instruments. Colors indicate soil type based on fi eld observations (Lin et al., 2006). Ten-minute precipitation amounts and event-based water samples were collected at the ridge top clearing (gray box) for isotope analysis. Soil water was collected biweekly at four nested lysimeter transects (orange circles) as described by Jin et al. (2011). Two unscreened groundwater wells were installed in the stream riparian zone to observe daily evolution of isotope composition (dark blue triangles). Depth to groundwater table was monitored in real-time in the stream riparian zone (teal triangle). Stream discharge and isotope composition values were monitored daily at the double V-notch weir located at the stream outlet (red triangle).


Stable Isotope MonitoringFrom 13 June 2008 to 16 Mar. 2012, a total of 4460 unfi ltered water samples were collected over the entire catchment to deter-mine the isotopic composition of various hydrologic pools and their linkages. Precipitation samples were collected 14 Jan. 2009–12 Feb. 2012 on an event basis using an automatic precipitation sampler (NSA 181/S model, Eigenbrodt GmbH & Co., Königsmoor, Ger-many) located in a clearing area on the ridge top (Fig. 1). To cap-ture intrastorm variability, precipitation was collected for the fi rst 30-min of a storm followed by three 6-hour intervals. Precipita-tion amount and isotope composition were assumed to be uniform over the entire catchment. Previous studies in central Pennsylvania have shown that one rain gauge per 12.6 km2 is suitable to resolve spatial variability of precipitation (Reich, 1966; O’Driscoll et al., 2005). As isotopic enrichment of summer precipitation associated with throughfall (0.11‰ and 2.9‰ for δ18O and δD, respectively; Pearce et al., 1986) is similar to analytical precision; thus the iso-topic composition of throughfall was assumed to be equal to the composition of the sampled precipitation.

Daily stream water and groundwater samples were collected using automatic samplers (2700 series, Teledyne Isco). Stream water was sampled directly upstream of the weir from March 2009 to March 2012. Groundwater was collected from March 2009 to March 2012 from the bottom of two screened wells (approximately 3 m deep), located in the stream riparian zone (Fig. 1).

As reported by Jin et al. (2011), soil water samples were collected approximately every 2 wk from four transects of nested suction-cup lysimeters (48-mm diameter, SoilMoisture 1900 series, Soil Moisture Inc., Santa Barbara, CA) emplaced in 2006 (Fig. 1, Table 1). Jin et al. (2011) only reported samples from the south planar hillslope of the catchment collected from 2006 to 2010. Th at same dataset is extended here through 2011 and we also include samples from the north side of the catchment and swale transects. Soil water sampling was oft en limited during the summer (June–Sep-tember) due to low soil water content and in the winter (Decem-ber–March) due to frozen soil condition and snow cover. Soil water samples were collected according to established USGS protocols (White et al., 2005) as described previously by Jin et al. (2011). Lysimeters were suctioned to 50 kPa 1 wk before sampling and thus only collected relatively mobile water within the soil. Th e four transects of lysimeters were located along a swale and a planar hill-slope each along the south and north slopes (Jin et al., 2010, 2011). Each transect contained lysimeter nests at three locations along the topographic gradient: ridge top, midslope, and valley fl oor. Th e midslope and valley fl oor of swales were dominated by the Rush-town soils while ridge tops and planar slopes were dominated by the Weikert soils (Fig. 1) (for soil chemistry and mineralogy, see Jin et al., 2010; Jin and Brantley, 2011; and Ma et al., 2011). Each nest consisted of 3 to 13 lysimeters (depending on soil thickness) installed at 10-cm interval depths and spaced 10 cm apart (Table

Table 1. Installed suction cup lysimeter depths and corresponding soil horizons for Transects 1–3, based on in situ observations (Lin, 2006). Th e soil horizons for Transect 4 were identifi ed by Jin et al. (2010). Organic Oe-horizon covers the upper ~5 cm of the entire catchment. Lysimeters located in the A-Horizon are colored green, B-Horizon yel-low, C-Horizon red, and in bedrock (R) brown. Locations with two lysimeters are labeled A and B aft er lysimeter depth.

Transect 1: North (south-facing) swale depression

Ridge top Midslope Valley fl oor

Soil series: Weikert Soil series: Rushtown Soil Series: Rushtown

Lysimeter depths

Soil horizon

Lysimeter depths

Soil horizon

Lysimeter depths

Soil horizon

cm cm cm

10 A 10 A 10 A

20 Bw 20 Bw2 20 Bw2

30 CR 40 BC 40 BC

40 R 60 BC 60 BC

80 C 80 C

100 C 100 C

140A/140B C 140 C

180 C 180 C

220 C 220 C

260 C 260 C

300 C 300 C

340 C 340 C

Transect 2: North (south-facing) planar hillslope

Soil series: Weikert Soil series: Weikert Soil series: Berks

10 A 10 A 10 Bw1

20 Bw 20 Bw 20 Bw2

30 CR 30 CR 30 Bw2

40 R 40A/40B R 40 Bw2

50A/50B R 50 Bw2

60 R 60 Bw3

160 R 70 C

220 R 80 C

100 C

120 C

Transect 3: South (north-facing) swale depression

Ridge top Midslope Valley fl oor

Soil series: Weikert Soil series: Rushtown Soil series: Rushtown

10 A 20 Bw2 10A/10B A

20 Bw 40 BC 20 Bw2

30 CR 60 BC 30 Bw3

80 C 40 BC

100 C 50 BC

120 C 60 BC

140 C 70 C

160 C 80 C

90 C


1). Th e fi rst lysimeter was installed at a depth of 10 cm with subse-quent lysimeters added until the point of augering refusal.

All liquid water samples were collected and stored in 30 mL amber glass or plastic vials with cone-lined caps to prevent evaporation. Sample bottles contained little or no headspace to minimize water-air exchange. Water samples were analyzed for water isotope com-position using laser absorption spectrometer (DT-100 Liquid Water Isotope Analyzer, Los Gatos Research, Inc., Mountain View, CA) following the IAEA Standard Operating Procedure (Newman et al., 2008). Th e analytical precision of the instrument is 0.1‰ and 0.8‰ for 18O and deuterium, respectively (Lis et al., 2008). Th e isotopic composition of the water was reported as a δ-value relative to Vienna Standard Mean Ocean Water (VSMOW):

æ ö÷ç ÷d = -ç ÷ç ÷çè øVSMOW(permil) 1 1 000X

XR

R, [1]

where R is the ratio of deuterium to hydrogen or 18O to 16O in the unknown (X) or VSMOW. Our quality assured isotopic data are available at the NSF-sponsored SSHCZO website (http://critical-zone.org/shale-hills, accessed 18 Apr. 2013).

Data AnalysesTo determine the isotope composition of precipitation for a given time period, the mean precipitation isotope composition was weighted by volume of precipitation and calculated using the fol-lowing equation:

1T

1

ni ii

nii

P

P=

=

dd =

åå

[2]

where δT is the amount-weighted precipitation value, Pi represents the amount (depth) of precipitation in an event, δi is the isotope composition of an event, and n is the number of events in the time period considered (week, month, or season). To determine temporal variations in isotope compositions and to account for the moder-ate sampling frequency of lysimeters, samples were averaged into four seasons: spring, summer, fall, and winter. Seasons were defi ned based on the astronomical calendar rather than the meteorological

calendar to more closely align with the growing season of the veg-etation in the catchment. For example, the spring season begins on the Spring Equinox and ends on the Summer Solstice, dates that vary from year to year. Th e growing season was defi ned as the com-bination of spring and summer while the nongrowing season was defi ned as the combination of fall and winter. Th is seasonal naming convention remains consistent throughout the study.

Violin plots (Hintze and Nelson, 1998) of the smoothed kernel density distributions (Bowman and Azzalini, 1997) were used to examine the isotopic distribution of seasonal precipitation inputs as well as the other hydrologic pools (soil water, stream, and groundwater). Th e size of the violin plot was normalized so the area of each plot was equal to one and the x axis scale was adjusted independently for each violin plot to allow for compari-sons of individual distribution shapes. Distributions were tested for normality using Shapiro–Wilk method (Shapiro and Wilk, 1965) at an a level of 0.05 as a measure of symmetry. For precipitation violin plots, amount-weighted concentrations were calculated as well as two weighted standard deviations. Observations were off set randomly along the x axis to allow for visualization. For the violin plots of various hydrologic pools, standard box and whisker plots were created. Boxes correspond to the inner-quartile and whiskers correspond to 1.5 of inner-quartile range. Violin plots of soil water were binned into three groups: shallow soil water (£30-cm depth), intermediate soil water (40–100-cm depth), and deep soil water (³120-cm depth). Th is was intended to facilitate the investigation of general soil water isotope composition patterns with respect to depth throughout the entire catchment with varying soil types. Statistical diff erences in medians of hydrologic pools of interest were tested using the Wilcoxon rank sum test, a nonparametric test, which does not assume normality and allows for unequal variance.

To determine the temporal evolution of soil water isotope composi-tion profi les and the associated uncertainty, the isotopic composi-tion of the soil water was spatially and temporally interpolated using an ordinary kriging method in the MATLAB soft ware (Th e MathWorks, Natick, MA). Soil water isotope compositions of lysimeters emplaced at equal depths were averaged for each collection date over all the sampling locations and applied to a depth-time grid with 1 cm spacing vertically and 5-d spacing horizontally. Th is was done to identify depths or times of either gradual or rapid changes in isotope composition throughout the catchment. Monthly amount-weighted isotope compositions of precipitation along with precipitation amounts were calculated for January 2009–December 2011 for reference. Th e color map thus created was applied to discern depth-time patterns of soil water iso-tope composition in relation to precipitation. Depth to the water table at one location in the stream riparian zone was fi ltered using a 10-d moving average to reduce noise and was superimposed on the above map to facilitate the investigation of its possible relation to soil water isotopes.

Transect 4: South (north-facing) planar hillslope

Soil series: Weikert Soil series: Weikert Soil series: Ernest

10 A 10 A 10 A

20 B 20 Bw 20 AE

30 C 30 Bw 30 Bt

40 C 40 Bt

50 C 60 2C


Spatial patterns of soil water isotope composition with depth and location in the catchment were created using stratified sampling of the measured data. Seasonal profiles were created to separate tem-poral patterns from spatial patterns. Seasonal groupings were per-formed following the convention as described above. As explained above, lysimeter samples were presented as catchment-averaged profiles, i.e., all sampled lysimeters at equivalent depths through-out the catchment for a given season were averaged to develop a composite averaged seasonal isotope profile. It was necessary to perform seasonal averaging of isotope profiles because such pro-files for any given sampling day were inherently noisy, making the analysis of general spatial individual local profiles and patterns dif-ficult. Determination of the seasonal standard deviation for each profile provided information regarding flow paths and residence times, with larger standard deviations typically associated with shorter residence times and smaller standard deviations typically associated with longer residence times. Isotope profiles were also stratified for each hillslope type (swale vs. planar hillslope), slope location (ridge top, midslope, and valley floor), and slope orienta-tion (south-facing vs. north-facing) by averaging soil water samples collected from equal depths at each location. In other words, for each season we calculated the average isotope composition and standard deviation for each sampling depth at each location. Because such profiles represent general seasonal patterns at given locations, depths with only one observation were removed from such profiles to limit misinterpretation. The winter and summer profiles were selected for comparison because the soils are typically the wettest during the winter and the driest during the summer months (Lin, 2006; Lin et al., 2006). Furthermore, the winter and summer isotope profiles make up the boundaries of the annual iso-tope profiles, providing the clearest differences in seasonal changes. In each soil water isotope profile, we have estimated representative soil horizons based on each lysimeter nest’s location and surveyed soil map and soil-bedrock interface (Table 1, Fig. 1). This may pro-vide additional insight into the influence of soil horizonation on hydrologic processes and thus on soil water isotope compositions.

6ResultsCatchment HydrologyAnnual precipitation amounts were 998 mm (2008), 1033 mm (2009), 896 mm (2010), and 1321 mm (2011) as recorded in the rain gauge. During the winters (December–March), the precipita-tion amounts (total water equivalent) were 251 mm (2008–2009), 351 mm (2009–2010), and 277 mm (2010–2011), consisting mostly of snow. Discharge ranged between 0 and 48 L s−1, with periods of zero discharge typically occurring between late June and early September. Large discharges (>20 L s−1) occurred immedi-ately after large storms or snowmelt events, after which discharge quickly decreased to <5 L s−1. Patterns of depth to water table were similar to discharge but had a slower rate of change. Rising water tables were associated with large storms and/or snowmelt events, but generally were found to recede more slowly than stream

discharge. Depth to water table within the stream riparian corridor ranged from 0.36 to 2.26 m during 2009–2011, with the deepest observed depth (well below the streambed) found at the end of the growing season (late August).

Precipitation Isotope CompositionFor the entire study period, the precipitation amount or depth-weighted average isotopic composition was -8.71‰ and -57.35‰ for δ18O and δD, respectively. For seasonally-averaged samples, the least depleted values were detected in the spring, with progressively more depleted values observed in the summer, fall, and then in the winter (Fig. 2). These data compare well to other observations of seasonal differences in precipitation isotope composition (DeWalle et al., 1997; O’Driscoll et al., 2005). This trend was attributed to the temperature dependent water fractionation associated with phase changes in the hydrologic cycle, including variable tempera-tures at the moisture source of water, variable temperatures at the precipitation site, and variations in evaporative fluxes throughout the year (Gat, 1996; Araguás-Araguás et al., 2000). The largest seasonal range in the precipitation’s isotopic signature occurred in the winter, while the smallest range was observed in the spring. The weighted kernel density distributions (Fig. 2) indicated that, for the full record, the isotopic composition of the precipitation was asymmetric and skewed toward more depleted values. The δ18O and δD distributions for the spring and the δ18O for the summer were relatively symmetrical and were considered normally distributed at the 0.05 a level using the Shapiro–Wilk test. The Wilcoxon rank sum test results showed that seasonal medians were statistically different between all seasons with the exception of the difference between spring and summer (p = 0.051 and p = 0.008 for δ18O and δD, respectively) and fall and winter (p = 0.059 and p = 0.010 for δ18O and δD, respectively), which were considered borderline at the 0.05% level.

Precipitation records were also analyzed to determine subseasonal temporal trends. Although the record contained substantial vari-ability, the time series of individual precipitation sample compo-sitions (Fig. 3) showed typical seasonal variations, with depleted values in the cold months as snow and subsequently less depleted values in the warm months as rain. Integration of precipitation samples to weekly amount-weighted isotope compositions (Fig. 4) preserved the strong seasonal trends, ranging from -25.80‰ to 1.31‰ for δ18O and from -200.72‰ to 6.47‰ for δD for the entire monitoring period. Further integration to monthly amount-weighted isotope compositions also preserved the seasonality but with much less variability (Fig. 5).

Two Local Meteoric Water Lines (LMWL) of the SSHCZO were created using all 331 six-hour precipitation samples, one follow-ing a standard least squares regression procedure and the other following a precipitation amount-weighted least squares regres-sion procedure (Hughes and Crawford, 2012). The precipitation amount-weighted least squares regression limits the influence of


individual precipitation samples with low amounts and possible evaporation biases which contribute little to the overall hydrology of the catchment. Th e slopes of both LMWLs at the SSHCZO closely matched the Global Meteoric Water Line proposed by Craig (1961) (Fig. 3), which is typical of the slightly more humid climate of central Pennsylvania.

Depth-Time Variability of Soil Water Isotope Compositi onsAggregation of soil water samples across the entire catchment showed clear patterns of isotope composition attenuation with depth (Fig. 4 and Fig. 6). Th e kernel density distribution of isotopic composition of all combined soil water samples was asymmetrical with values

Fig. 3. A total of 331 precipitation samples collected for isotope composition from 14 Jan. 2009 to 12 Feb. 2012. Precipitation is shown as both rain (gray circles) and snow (blue stars). All samples were collected over 6-h intervals. Th e unweighted local meteoric water line (LMWL, solid line) and amount-weighted local meteoric water line (black dashed line) were calculated using all precipitation samples (a). Th e global meteoric water line (GMWL, gray dashed line) falls behind the unweighted LMWL. Th e time series of δ18O for all precipitation samples and the daily precipitation record (b) with precipitation grouped into rain and snow.

Fig. 2. Violin plots of the distribution of 18O (a) and deuterium (b) compositions in precipitation. Th e plot for the total record consists of precipita-tion events collected from 14 Jan. 2009 to 12 Feb. 2012, with sample number n indicated. Seasons are divided based on astronomical calendar to more closely correspond to the growing season within the catchment. Colors of the seasons correspond to colors used in seasonal soil water profi le fi gures (Fig. 3, Fig. 7, Fig. 8, and Fig. 9). Th e shape of the violin plot is the normalized, weighted kernel density distribution, with each x axis scaled indepen-dently to allow for comparisons of distribution shapes. Plots with shaded backgrounds correspond to normal distributions at the 0.05 a level based on Shapiro–Wilk test for normality. Th e center square is the amount-weighted mean composition and the bars are plus and minus two weighted-standard deviations. Black points are individual data points, jittered horizontally for visualization.


skewed toward more depleted compositions (Fig. 6). With further separation into shallow (£30 cm), intermediate (40–100 cm), and deep soil water (³120 cm), the kernel density distributions remained asymmetrical. Th e stream water kernel density distribution was

slightly bimodal with values skewed toward less depleted composi-tions. Groundwater kernel density distributions were tightly grouped but not normally distributed at the 0.05% level. Th e bimodal δD dis-tribution in the groundwater was due to the two monitoring wells

Fig. 4. Temporal variation in weekly precipitation amount (dark bars), and the isotope composition of precipitation (gray bars) and soil water at three depths (shallow, intermediate, and deep). Soil water data before November 2008 are not shown since precipitation record did not begin until January 2009. Shaded regions correspond to one standard deviation of each soil depth. Data for each soil depth is averaged over the entire catchment. Ground-water mean (green line) is shown as a reference.

Fig. 5. Monthly amount-weighted isotope composition in precipitation, the kriged map of soil water isotope composition through depth and time, and monthly precipitation amount. Red values correspond to less-depleted isotope composition while blue values correspond to more-depleted values. Black circles correspond to catchment-averaged lysimeters (see text). Ten-day moving averages of the groundwater level in the valley fl oor are plotted as a black line. Kriging uncertainty contours are plotted over the same space-time domain (see supplemental materials). Ordinary kriging and kriging variance were computed in Matlab.


having slightly diff erent median isotopic compositions (Fig. 6). As observed by Jin et al. (2011) using the smaller dataset, it was apparent that the range of isotope compositions substantially decreased from precipitation to soil water and to stream water and to groundwater, a trend which was preserved in this larger dataset. During the entirety of this study, the groundwater isotope composition remained near its average of -8.82‰ and -56.53‰ for δ18O and δD, respectively.

Isotope compositions of soil water varied seasonally, closely refl ect-ing the precipitation composition (Fig. 4). Generally, more depleted compositions of soil water were found in the winter and spring and less depleted compositions in the summer and fall. However, for a given year, the soil water sampled during the spring (rather than winter) was generally the most depleted due to the snowmelt. Com-positions became less depleted through the spring and summer months, with soil water isotope compositions becoming the least

depleted in the fall; however, the degree of such an enrichment trend varied from year to year (Fig. 4).

Seasonal diff erences of soil water isotope composition were also refl ected in catchment-wide average vertical profi les, with δ18O and δD exhibiting strong attenuation with soil depth (Fig. 7). In these profi les, the shallow soil waters (£30-cm depth) showed larger standard deviations and had a mean similar to the isotope composition of seasonal precipitation average. Th is seasonal ampli-tude was attenuated with depth as soil water isotope compositions approached the groundwater average. Overall, the winter isotope composition profi le was more depleted than the groundwater aver-age while the summer isotope composition profi le was less depleted. Furthermore, the winter profi le had a larger standard deviation for shallower soil depths (<100 cm) when compared to the summer. Th e Wilcoxon rank sum tests showed winter and summer profi les in the shallow (£30 cm) and intermediate (40–100 cm) depths had

Fig. 6. Violin plots of the distribution of deuterium (a) and 18O (b) composition for each hydrologic pool within the catchment sampled from 13 June 2008 to 16 Mar. 2012, with sample number, n, indicated. Th e ratio of each hydrologic pool’s variance to the weighted precipitation variance is shown in bottom corner of each panel. Th e full record was split into four seasons, spring (c), summer (d), fall (e), and winter (f ), based on astronomical calendar. Th e full soil profi le violin plot (orange) is a composite of shallow soil water (purple, £30 cm), intermediate soil water (teal, 40–100 cm), and deep soil water (blue, ³120 cm). Th e shape of the violin plot is the normalized kernel density distribution, with each x axis scaled independently to allow for comparisons of distribution shapes. Plots with shaded backgrounds correspond to normal distributions at the 0.05 a level based on Shapiro–Wilk test for normality. Inside each violin plot is a standard bar plot. Th e white point corresponds to the median of the data, inner black box is interquartile range, and whiskers cover 1.5 interquartile range. Outliers are left off for clarity, but can be seen in the tails of the kernel density distribution.


statistically diff erent medians (p = 2.89 ´ 10-12 and p = 5.61 ´10-10, respectively), while medians for winter and summer profi les at deep depths (³120 cm) were statistically indistinguishable (p = 0.61). Th e stable isotope composition profi les for the transitional seasons of fall and spring were intermediate between that of the wet and dry seasons with shallow soil water closely matching seasonally-weighted precipitation compositions (Fig. 7). Both these profi les had relatively constant means and decreasing standard deviations with depth. Th e isotope profi le during the spring season had rela-tively large standard deviation that was indicative of wetter soils and shorter residence times of water, whereas the fall season had less variation indicative of drier soils and longer residence times. Th ese trends of seasonally dependent soil moisture matched other fi eld observations (Lin, 2006; Lin et al., 2006). Both isotope profi les approached the groundwater average at deeper soil depths. Interest-ingly, only the spring profi le was more depleted than the associated seasonal precipitation composition for the entire soil depth.

Closer investigation using weekly and monthly time series revealed more details of the diff erences in soil water isotope composition evolution with depth. Water sampled from the lysimeters located in the shallow soils (£30 cm) had the largest seasonality while water

sampled from the deep lysimeters (³120 cm) showed little seasonal-ity (Fig. 4). Th e deepest soil water isotope compositions were rela-tively constant through time and space and closely matched the groundwater isotope composition (Fig. 5). Th is is likely due to the fact that these deep soil waters were collected in the midslope and valley fl oor where soil waters are known to have been derived from not only precipitation inputs but also soil water from upslope and groundwater inputs (Jin et al., 2011). Soil water sampled from the upper 30 cm had the largest δ18O variations, while maximum and minimum isotope compositions deeper in the soil profi les typically occurred later in time indicating a phase lag. Th e time delay in the amplitude of the isotope soil profi les is especially evident in the upper soil layers and can be seen by comparing isotope composition of precipitation and soil water sampled at 10 cm (Fig. 5).

Interannual variability is also apparent through closer investiga-tion of higher temporal resolution time series. For instance, during the spring and summer of 2009 and 2011, the soil water isotope composition became less depleted than the groundwater average, with the shallow soil water showing the strongest trend (Fig. 4 and Fig. 5). Total enrichment of shallow soil water (£30 cm) from April to October was approximately 3‰ for δ18O. Enrichment of intermediate and deep soils also occurred but was less prominent.

During the spring and summer of 2010, the soil water isotope composition showed a slightly diff erent attenuation with depth as compared to that in 2009 or 2011 (Fig. 4 and Fig. 5). Over that period, soil water was more depleted than the groundwater average. Th is highlights that interannual variability in isotope composition cannot be neglected. Th roughout the remainder of the spring and summer in 2010, the soil water displayed a similar enrichment trend as in 2009 and 2011 but with a larger magnitude of 5‰ for δ18O in the shallow soil layer.

Spati al Variability of Soil Water Isotope Compositi onsNorth and South SlopesTo determine spatial patterns in the soil water stable isotope compo-sition, variations of soil isotope profi les on the north (south-facing) and south (north-facing) slopes were compared during the winter and summer. For both the north and south slopes, the winter pro-fi les were depleted while the summer profi les were enriched (similar to precipitation). Th e seasonal variation was attenuated with depth and approached the groundwater average below ~100 cm (Fig. 8). However, the amplitude of the winter (wet season) profi les diff ered between the two slopes, with a more depleted signature observed on the north slope in the upper 20 cm as compared to the south slope. Furthermore, the diff erence between isotope compositions in the upper layer of the winter and summer profi les was larger for the north slope as compared to the south slope. In addition, for both the north and south slopes, the winter profi le had a larger standard deviation (amplitude) than the summer profi le. Th e winter profi le

Fig. 7. Seasonal δ18O profi les of soil water for wet (winter) and dry (summer) seasons (a) and transitional seasons (spring and fall) (b) based on the 3-yr monitoring. Isotope composition at each depth was averaged over the entire catchment. One standard deviation is represented with shaded regions. Depths with only one sample were removed for clarity. For reference, groundwater mean and one stan-dard deviation is shown as a gray bar, and the amount weighted precip-itation compositions with weighted standard error for corresponding seasons are shown as colored arrows at the top.


for both the north and south profi les showed considerable variabil-ity down to a depth of ~80 cm while the summer profi les showed less variability and were relatively smooth.

Ridge Top, Midslope, and Valley FloorSeasonal soil water isotopic composition profi les also diff ered with slope location regardless of the north or south slopes (ridge top, midslope, and valley fl oor). Th e amplitude of the winter profi le for each slope location showed larger variability compared to the summer profi les (Fig. 9). Profi les on the ridge top were limited by soil depth (<50 cm) and generally refl ected winter and summer precipitation. Th e winter profi le at midslope showed increases in amplitude variation at both 20- and 60-cm depths where horizon interfaces are located (Table 1), layers identifi ed by Jin et al. (2011) as high-fl ow zones. Th e winter profi le in the valley fl oor showed the greatest variation in the amplitude in the upper part of the profi le.

Swale and Planar HillslopesFor both the winter and summer profi les in the swale depressions, there were specific depths where the standard deviation of iso-topic compositions increased (Fig. 10). Specifi cally, in the winter,

the standard deviations increased at 30- and 60-cm depths. Th ese depths, which correspond to the soil-bedrock interface on the ridge top and the B–C horizon interface at the midslope and valley fl oor lysimeter nests (Table 1). Additionally, the swale summer profi le showed increases at depths of 60 and 100 cm. Th e planar hillslope isotope profi les, on the other hand, were relatively more uniform, but with large variation in amplitude in at various depths during the wet season. At the deepest depths, both swale and planar profi les were relatively constant approaching the groundwater mean (Fig. 10).

6DiscussionAtt enuati on of Seasonality with Soil DepthTh e seasonal attenuation with depth is characteristic of advective-dispersive transport and mixing (Leibundgut et al., 2009) and has been observed in previous studies (Eichinger et al., 1984; Darling and Bath, 1988; Geake and Foster, 1989; Clark and Fritz, 1997; Tang and Feng, 2001; Lee et al., 2007) including work at the SSHCZO (Jin et al., 2011). However, the degree and cause of attenuation reported in these earlier studies diff ered. Leibundgut et al. (2009) pointed out that the depth of seasonal attenuation decreases as soils become fi ner grained. For example, in deep gravel profi les of western Germany, Eichinger et al. (1984) found the sea-sonal isotopic profi le was attenuated over the fi rst 9 m; while in the English chalk, heavy isotopic attenuation was found in the fi rst meter (Darling and Bath, 1988; Geake and Foster, 1989). At the SSHCZO, the silt loam or silty-clay loam apparently provided strong attenuation of the seasonal isotopic profi le typically within the upper ~1 m. One explanation for the relatively rapid attenua-tion of the seasonal isotope profi les is hydrodynamic mixing during percolation through permeable soils, a phenomenon observed in the majority of soil profi les at the SSHCZO (Fig. 6–10).

Evidence for Preferenti al FlowPrevious work at the SSHCZO has identifi ed the importance of both vertical and lateral preferential fl ow in the catchment (Lin, 2006; Lin and Zhou, 2008; Graham and Lin, 2011; Jin et al., 2011). Vertical preferential fl ow has been identifi ed as occurring due to various macropores, hydrophobicity, and high-permeability structured soils, while lateral preferential fl ow occurs along lateral macropores and soil horizon and/or soil-bedrock interfaces (Lin and Zhou, 2008; Graham and Lin, 2011; Jin et al., 2011). Our fi ndings in this study support these observations and provide fur-ther insight into specifi c locations within soil profi les where more preferential fl ow has been more active.

Soil water isotopic profi les help to understand the timing and direction of water movement at the SSHCZO. In the summer profi les, it was clear that the standard deviation of soil water iso-topic compositions remained relatively low throughout the entire soil profi le down to 90 cm (Fig. 7). Th is is consistent with longer residence times and smaller change within the soil profi le. Th is pattern changed slightly in the deepest soils: at a depth of 100

Fig. 8. Seasonally-averaged δ18O composition in soil profi les for the north (south-facing) (a) and south (north-facing) (b) slopes based on the 3-yr monitoring. One standard deviation is represented with shaded regions. Depths with only one sample were removed for clar-ity. Typical soil horizon interfaces are shown as black dashed lines. Groundwater mean and one standard deviation are shown as gray bars for reference.


cm, the standard deviation expanded. To explain this increase in standard deviation, water with a diff erent isotope composition must have replaced the water in deeper soils without displac-ing the water in the shallower soil layers. Th ere are two possible explanations for this phenomenon when observed near the valley fl oor: (i) downward preferential fl ow through macropores located either in the valley itself or derived from interfl ow from upslope (that itself derived from water inputs through macropores located further upslope), or (ii) upward groundwater movement through either capillary rise and/or a rising water table. Upward ground-water movement was unlikely since the isotope composition of groundwater remains nearly constant throughout the summer months and thus would not explain the increased standard deviation. Th is means that preferential fl ow that bypassed the shallow part of the soil profi le (either through macropores in the valley fl oor or from the upslope area) was likely the cause of this increased standard deviation. Furthermore, similar pat-terns for each landform type and hillslope location during the summer months were found, suggesting vertical preferential fl ow bypassed the shallow soils in other locations in the catch-ment (Fig. 8, Fig. 9, and Fig. 10). Th is appearance during only the summer months was likely due to the temporary existence of shrinkage cracks and hydrophobicity in forested soils caused by low soil moisture conditions.

Soil water isotopic composition profi les also provided evidence of lateral preferential fl ow within the SSHCZO. For example, the winter isotope profi le for soil water measured at midslope showed increases in standard deviation at depths of 20 and 60 cm associ-ated with the A–B and B–C horizon interfaces (Table 1, Fig. 9). Th is increased variation in isotopic composition was likely caused by relatively rapid exchange of water of diff ering isotopic compo-sition at locations of high fl ow. By comparing depths with high fl ows to soil horizon depths, it is possible to infer preferential fl ow along the soil horizon or soil-bedrock interfaces. Th is pattern of increased standard deviation at depths related to soil horizon or soil-bedrock interfaces occurred in other soil profi les during the wet season (including the valley fl oor, swale, and planar hillslopes), suggesting that lateral preferential fl ow was widespread within this catchment. Note that evidence for such lateral preferential fl ow was not observed during the summer months when soils were dry (Fig. 8–10). Th is general model for fast water fl ow through ver-tical macropores and horizontal soil interfaces was also invoked by Jin et al. (2011) to explain the water isotopic data and water chemistry data specifi cally for the south planar hillslope. Addition-ally, the larger standard deviations throughout the isotope pro-fi les during the winter months suggest that the soil profi les were more hydrologically active and residence times were shorter in the winter than during the summer months. All these preferential fl ow

Fig. 9. Seasonally-averaged δ18O composition in soil profi les for each slope location based on the 3-yr monitoring: ridge top (a), midslope (b), and valley fl oor (c). One standard deviation is represented with shaded regions. Depths with only one sample were removed for clarity. Typical soil horizon interfaces are shown for each slope location. Groundwater mean and one standard deviation are shown as gray bars for reference.


observations from soil water isotopic compositions are consistent with the results reported by Lin (2006) and Lin et al. (2006).

Snowmelt DynamicsSoil water isotope composition profiles also provided insight into snowmelt dynamics, which play an important role in solute transport and chemical weathering diff erences between the slopes at the SSHCZO. Th is was best observed when comparing the profi les from the north and south slopes: the north slope had a more depleted composition in the uppermost horizons during the winter months as compared to the south slope (Fig. 8). Th is more depleted signature was due to varying snowmelt dynamics. Due to the small size of the catchment, the amount and isotope composition of snowfall is assumed to be the same for both the slopes. Th e east–west orientation of the catchment, however, pro-duces an asymmetry in solar radiation between the two diff erent sun-facing slopes. Th e higher solar radiation on the north slope supports a relatively thin snowpack (typically <30 cm) with likely earlier melting and infi ltration into the surface soils. Th e infi ltra-tion of isotopically-depleted snowmelt thus altered the composi-tion of the upper 20 cm of the soil, resulting in a soil water isotope composition of -16‰ for δ18O on the north slope. Conversely, on the south slope, the snowpack remained intact for longer periods

in the winter months where sublimation or evaporation could enrich the isotopic composition of the residual snowpack. Th is phenomena has been observed previously in other catchments (Moser and Stichler, 1974; Whillans and Grootes, 1985). Once the snowpack on the south slope melted and infi ltrated the soil, this less depleted signature would then be imparted to the soil profi le. Another potential explanation of the diff erences in winter isotopic profi les for the north and south slope is the possible redistribution of snowmelt throughout a larger depth on the south slope. Th ese diff erences in water dynamics on the south vs. the north have been invoked as partial explanations for the observed diff erences in the extent of weathering on the two sides [e.g., the soils on the south side are more chemically depleted in many elements than the north side (Ma et al., 2013)].

Th e temporal evolution of soil water isotopes with depth provided other insights into the importance of snow in soil water dynamics at the SSHCZO. For example, aft er the winter of 2010, an extreme winter that produced that produced 33% more precipitation, the spring and summer isotope profi les showed dramatically more isotopic depletion throughout the entire profi les (Fig. 5). Aft er the fi rst sampling in the spring, relatively depleted isotope com-positions appeared in depths from 10 to >150 cm. Th is depleted signature was due to isotopically-depleted wintertime precipitation, melting and infi ltration, which subsequently mixed intermittently with deeper soil (150–175 cm) and groundwater. Th is depleted isotope signature remained in soil water through mid May 2010, when spring and summer precipitation gradually replaced the shallow soil water while deeper soils (>25 cm) continued to hold water with the depleted signature until fall. During this period, most soil profi les showed high moisture content during the spring snowmelt and this moisture was retained through early summer. As the growing season progressed, transpiration gradually removed the matrix water, which then began to absorb spring and summer-time precipitation. Still, the soil water at the intermediate depths retained the winter precipitation signature through the entire growing season, showing somewhat depleted signatures at 70-cm depth in October 2010 (Fig. 5).

Evapotranspirati on DynamicsSeasonal enrichments of soil water isotope compositions were found to occur during each growing season (spring–summer) (Fig. 4 and 5). Th ere are two factors contributing to this enrichment trend: (i) evaporation of water from the soil or (ii) input of less depleted summer precipitation. To determine the degree of grow-ing season evaporation, soil water samples were plotted against the unweighted and weighted LMWL (Fig. 11). Variations along the LMWL correspond to the seasonal precipitation composi-tions while deviations to the right of the line are indicative of evaporative eff ects due to diff erences in fractionation rates of 18O and deuterium (Barnes and Allison, 1988), allowing for distinc-tion of enrichment trends. During the summer months, all the soil water samples fell on or above the unweighted LMWL and

Fig. 10. Seasonally-averaged δ18O composition in soil profi les for hill-slope type based on 3-yr monitoring: swale depression (a) and pla-nar hillslope (b). One standard deviation is represented with shaded regions. Depths with only one sample were removed for clarity. Typical soil horizon interfaces are shown for each hillslope type. Groundwater mean and one standard deviation are shown as gray bars for reference.


near the weighted LMWL, with samples from shallow soils (£30 cm) near the spring and summer precipitation compositions and samples from intermediate soils (40–100 cm) near the groundwa-ter average (Fig. 11), indicating mobile soil water showed little to no evaporative signal, but was highly dependent on the precipita-tion isotope composition. Additionally, investigation of deep soil waters (³120 cm) showed no evidence of evaporation (data not shown). Preliminary work suggests soil water samples fell above the unweighted LMWL due to evaporative biases introduced from high-resolution sampling of precipitation. Additional work is now underway to validate these fi ndings and to establish a more robust LMWL for hydrological applications. Because of the limited role of evaporation, transpiration is believed to be the main contribu-tor of the evapotranspiration fl ux in this dense, forested catch-ment. However, the lysimeters may not have adequately sampled tightly-bound water in the soils and currently a study is underway to assess this eff ect.

6Summary and ConclusionsTh e stable isotope network developed at the SSHCZO has pro-vided extensive monitoring of the evolution and patterns of water stable isotopes for over 3 yr, beginning in 2008. The isotope compositions of the hydrologic pools most likely to infl uence soil profi les were established through high frequency sampling of pre-cipitation and groundwater. Comprehensive investigations of soil water isotope spatiotemporal patterns reported here have provided the following conclusions:

1. Shallow soil water was highly infl uenced by seasonal pre-cipitation input composition. Monitored soil water clearly refl ected the seasonal isotopic signal of precipitation, with more depleted values typically occurring in the winter months and less depleted values occurring in the summer months (Fig. 4).

2. Th e seasonal amplitude of soil water isotope composition was attenuated within depths of 1–2 m, approaching a nearly constant value similar to the groundwater composition (Fig. 7). Th is successive attenuation suggests the importance of hydrodynamic mixing, which integrates multiple seasons of precipitation.

3. Large variations in the winter soil water isotope values at dif-ferent depths in the profi les were observed at most sites and we interpreted this variation as evidence for preferential fl ow (Fig. 8–10). Increased variations were likely due to intermit-tent snowmelt, the eff ects of soil structure and horizons on macropore fl ow, and low residence times. Relatively smooth isotope profi les were observed in the summer season, indi-cating preferential fl ow is less frequent or more diffi cult to observe during the summer.

4. Diff erences in winter soil water isotopic profi les between the north (south-facing) and south (north-facing) slopes appear to be explained by asymmetric solar radiation balance, inducing earlier snowmelt and infi ltration on the north slope, but subli-mation and snowpack enrichment on the south slope (Fig. 8).

5. Extreme winter precipitation amounts during 2009–2010 infl uenced late growing season soil water isotope composi-tions (Fig. 5). Tracking of the depleted signatures through depth and time suggested relatively static hydrologic

Fig. 11. Shallow (£30 cm depth) and intermediate (40–100 cm depth) soil water sampled during the summer (dry) season with Susquehanna-Shale Hills Critical Zone Observatory (SSHCZO) Local Meteoric Water Line. Amount weighted precipitation compositions with weighted standard error bars for each of the four seasons are shown for reference. Green region indicates groundwater average ( ± one standard deviation) for each isotope composition.


conditions where residence times in the deep vadose zone’s mobile domain become longer through the growing season.

6. Summer soil water profiles aggregated across the catchment showed little evaporative effect (Fig. 7 and 11). The soil water isotope composition at the SSHCZO is dominated by the precipitation input, with minor evaporative enrichment, and this is consistent with transpiration dominating evapotrans-piration in this dense, temperate forest environment.

One limitation of this study is that soil water sampled from suc-tion-cup lysimeters does not reflect the most tightly-held water within the soil matrix (Darling and Bath, 1988; Landon et al., 1999; Figueroa-Johnson et al., 2007; Brooks et al., 2009). Cur-rently, new experiments are underway in the SSHCZO to resolve the mobile and less-mobile soil water fractions. For example, stud-ies are underway by the SSHCZO ecology team to investigate tree species water usage for addressing interception, transpired water, and root uptake. By monitoring tree xylem and root δ18O compo-sition, we will be able to determine effective depth of root water extraction. Additional work is underway to synthesize isotopic data into the greater SSHCZO conceptual hydrologic model and to numerically model isotopic age at the SSHCZO. Lastly, recent work has identified the importance of fractured bedrock and deep groundwater as important water sources to headwater streams (Burton et al., 2002). Drilling to explore deep vadose zone and groundwater at the SSHCZO is ongoing.

AcknowledgmentsWe thank those who installed the soil lysimeters, B. Ketchum. We also thank those who helped with various fieldwork, including B. Ketchum, R. Ravella, J. Williams, E. Herndon, M. Holleran, Z. Ruge, T. Yesavage, L. Liermann, K. Bazilevskaya, D. Andrews, C. Duffy, R. Jones, and L. Eccles. Financial support was provided by the National Science Foundation under Grant No. EAR- 0725019, 2008-2013 for Susquehanna-Shale Hills Critical Zone Observatory to CJD (PI).

ReferencesAndrews, D.M., H. Lin, Q. Zhu, L. Jin, and S.L. Brantley. 2011. Hot spots and hot

moments of dissolved organic carbon export and soil organic carbon stor-age in the Shale Hills Catchment. Vadose Zone J. 10:943–954. doi:10.2136/vzj2010.0149

Araguás-Araguás, L., K. Froehlich, and K. Rozanski. 2000. Deuterium and oxygen-18 isotope composition of precipitation and atmospheric moisture. Hydrol. Processes 14(8):1341–1355. doi:10.1002/1099-1085(20000615)14:8<1341::AID-HYP983>3.0.CO;2-Z

Barnes, C.J., and G.B. Allison. 1983. The distribution of deuterium and 18 O in dry soils: 1. Theory. J. Hydrol. 60:141–156. doi:10.1016/0022-1694(83)90018-5

Barnes, C., and G. Allison. 1988. Tracing of water movement in the unsaturated zone using stable isotopes of hydrogen and oxygen. J. Hydrol. 100:143–176. doi:10.1016/0022-1694(88)90184-9

Bowman, A.W., and Azzalini, A. 1997. Applied smoothing techniques for data analysis. Oxford Univ. Press, New York.

Brooks, J.R., H.R. Barnard, R. Coulombe, and J.J. McDonnell. 2009. Ecohydro-logic separation of water between trees and streams in a Mediterranean climate. Nat. Geosci. 3(2):100–104. doi:10.1038/ngeo722

Burton, W.C., L.N. Plummer, E. Busenberg, B.D. Lindsey, and W.J. Gburek. 2002. Influence of fracture anisotropy on ground water ages and chemistry, Valley and Ridge Province, Pennsylvania. Ground Water 40(3):242–257. doi:10.1111/j.1745-6584.2002.tb02652.x

Clark, I., and P. Fritz. 1997. Environmental isotopes in hydrogeology. Lewis Pub-lishers, Boca Raton, FL.

Craig, H. 1961. Isotopic variations in meteoric waters. Science 133(3465):1702–1703. doi:10.1126/science.133.3465.1702

Criss, R.E., and M.L. Davisson. 1996. Isotopic imaging of surface water/ground-water interactions, Sacramento Valley, California. J. Hydrol. 178(1–4):205–222. doi:10.1016/0022-1694(96)83733-4

Darling, W.G., and H. Bath. 1988. A stable isotope study of recharge pro-cesses in the English Chalk. J. Hydrol. 101(1–4):31–46. doi:10.1016/0022-1694(88)90026-1

Davisson, M., and R. Criss. 1993. Stable isotope imaging of a dynamic ground-water system in the southwestern Sacramento Valley, California, USA. J. Hy-drol. 144:213–246. doi:10.1016/0022-1694(93)90173-7

DeWalle, D.R., P.J. Edwards, B.R. Swistock, R. Aravena, and R.J. Drimmie. 1997. Seasonal isotope hydrology of three Appalachian forest catch-ments. Hydrol. Processes 11:1895–1906. doi:10.1002/(SICI)1099-1085(199712)11:15<1895::AID-HYP538>3.0.CO;2-#

Duffy, C. 1996. A two-state integral-balance model for soil moisture and groundwater dynamics in complex terrain. Water Resour. Res. 32(8):2421–2434. doi:10.1029/96WR01049

Ehleringer, J.R., and T.E. Dawson. 1992. Water uptake by plants: Perspectives from stable isotope composition. Plant Cell Environ. 15(9):1073–1082. doi:10.1111/j.1365-3040.1992.tb01657.x

Eichinger, L., B. Merkel, G. Nemeth, J. Salvamoser, and W. Stichler. 1984. Seep-age and velocity determinations in unsaturated quaternary gravel. In P. Ud-luft et al., editors, Recent investigations in the Zone of Areation. Technical University of Munich, Munich, West-Germany. p. 303–313.

Figueroa-Johnson, M.A., J.A. Tindall, and M. Friedel. 2007. A comparison of 18Oδ composition of water extracted from suction lysimeters, centrifu-gation, and azeotropic distillation. Water Air Soil Pollut. 184(1–4):63–75. doi:10.1007/s11270-007-9399-8

Gat, J.R. 1971. Comments on the stable isotope method in regional ground-water investigations. Water Resour. Res. 7(4):980–993. doi:10.1029/WR007i004p00980

Gat, J.R. 1996. Oxygen and hydrogen isotopes in the hydrologic cycle. Annu. Rev. Earth Planet. Sci. 24(1):225–262. doi:10.1146/annurev.earth.24.1.225

Gazis, C., and X. Feng. 2004. A stable isotope study of soil water: Evidence for mixing and preferential flow paths. Geoderma 119(1–2):97–111. doi:10.1016/S0016-7061(03)00243-X

Geake, A., and S. Foster. 1989. Sequential isotope and solute profiling in the unsaturated zone of British Chalk. Hydrol. Sci. J. 34(1):79–96. doi:10.1080/02626668909491310

Gierke, C. 2012. Sourcing tree water in the Sacramento Mountains of New Mexico: A stable isotope study. M.S. diss. New Mexico Institute of Mining and Technology, Socorro.

Graham, C.B., and H.S. Lin. 2011. Controls and frequency of preferential flow occurrence: A 175-event analysis. Vadose Zone J. 10(3):816–831. doi:10.2136/vzj2010.0119

Hintze, J., and R. Nelson. 1998. Violin plots: A box plot-density trace synergism. Am. Stat. 52(2):37–41.

Holmes, G.H. 2011. Using δ2H and δ18O to Determine the flowpath and tim-escales of water at the Susquehanna Shale Hills Critical Zone Observatory. M.S. diss. Pennsylvania State University, University Park.

Hughes, C.E., and J. Crawford. 2012. A new precipitation weighted method for determining the meteoric water line for hydrological applications demon-strated using Australian and global GNIP data. J. Hydrol. 464–465:344–351. doi:10.1016/j.jhydrol.2012.07.029

Jin, L., D.M. Andrews, G.H. Holmes, H. Lin, and S.L. Brantley. 2011. Opening the “black box”: Water chemistry reveals hydrological controls on weathering in the Susquehanna Shale Hills Critical Zone Observatory. Vadose Zone J. 10(3):928–942. doi:10.2136/vzj2010.0133

Jin, L., and S.L. Brantley. 2011. Soil chemistry and shale weathering on a hill-slope influenced by convergent hydrologic flow regime at the Susque-hanna/Shale Hills Critical Zone Observatory. Appl. Geochem. 26:S51–S56. doi:10.1016/j.apgeochem.2011.03.027

Jin, L., R. Ravella, B. Ketchum, P.R. Bierman, P. Heaney, T. White, and S.L. Brant-ley. 2010. Mineral weathering and elemental transport during hillslope evo-lution at the Susquehanna/Shale Hills Critical Zone Observatory. Geochim. Cosmochim. Acta 74(13):3669–3691. doi:10.1016/j.gca.2010.03.036

Kirchner, J.W. 2003. A double paradox in catchment hydrology and geochemis-try. Hydrol. Processes 17(4):871–874. doi:10.1002/hyp.5108

Kuntz, B.W., S. Rubin, B. Berkowitz, and K. Singha. 2011. Quantifying sol-ute transport at the Shale Hills Critical Zone Observatory. Vadose Zone J. 10(3):843–857. doi:10.2136/vzj2010.0130

Landon, M., G. Delin, S. Komor, and C. Regan. 1999. Comparison of the stable-isotopic composition of soil water collected from suction lysimeters, wick samplers, and cores in a sandy unsaturated zone. J. Hydrol. 224(1–2):45–54. doi:10.1016/S0022-1694(99)00120-1

Leaney, F., K. Smettem, and D. Chittleborough. 1993. Estimating the contribu-tion of preferential flow to subsurface runoff from a hillslope using deuteri-um and chloride. J. Hydrol. 147:83–103. doi:10.1016/0022-1694(93)90076-L

Lee, K.-S., J.-M. Kim, D.-R. Lee, Y. Kim, and D. Lee. 2007. Analysis of water move-ment through an unsaturated soil zone in Jeju Island, Korea using stable oxygen and hydrogen isotopes. J. Hydrol. 345(3–4):199–211. doi:10.1016/j.jhydrol.2007.08.006


Leibundgut, C., P. Maloszewski, and C. Külls. 2009. Environmental tracers. In , C. Leibundgut et al., editors, Tracers in hydrology. Wiley, New York. p. 13–56..

Lin, H. 2006. Temporal stability of soil moisture spatial pattern and subsurface preferential flow pathways in the Shale Hills Catchment. Vadose Zone J. 5(1):317–340. doi:10.2136/vzj2005.0058

Lin, H.S., W. Kogelmann, C. Walker, and M.A. Bruns. 2006. Soil moisture pat-terns in a forested catchment: A hydropedological perspective. Geoderma 131(3–4):345–368. doi:10.1016/j.geoderma.2005.03.013

Lin, H., and X. Zhou. 2008. Evidence of subsurface preferential flow using soil hydrologic monitoring in the Shale Hills catchment. Eur. J. Soil Sci. 59(1):34–49. doi:10.1111/j.1365-2389.2007.00988.x

Lis, G., L.I. Wassenaar, and M.J. Hendry. 2008. High-precision laser spectros-copy D/H and 18O/16O measurements of microliter natural water samples. Anal. Chem. 80(1):287–293. doi:10.1021/ac701716q

Lynch, J. 1976. Effects of antecedent soil moisture on storm hydrographs. Ph.D. diss., Pennsylvania State University, University Park.

Ma, L., F. Chabaux, N. West, E. Kirby, L. Jin, and S. Brantley. 2013. Regolith pro-duction and transport in the Susquehanna Shale Hills Critical Zone Obser-vatory, Part 1: Insights from U-series isotopes. J. Geophys. Res.: Earth Surf. (in press).

Ma, L., L. Jin, and S.L. Brantley. 2011. How mineralogy and slope aspect af-fect REE release and fractionation during shale weathering in the Susque-hanna/Shale Hills Critical Zone Observatory. Chem. Geol. 290(1–2):31–49. doi:10.1016/j.chemgeo.2011.08.013

McDonnell, J.J. 1990. A Rationale for old water discharge through macro-pores in a steep, humid catchment. Water Resour. Res. 26(11):2821–2832. doi:10.1029/WR026i011p02821

Moser, H., and W. Stichler. 1974. Deuterium and Oxygen-18 contents as an in-dex of the properties of snow covers. International Association of Hydro-logical Sciences Publication 144. IAHS, Potsdam, Germany. p.122–135.

National Oceanographic and Atmospheric Administration. 2007. U.S. divisional and station climatic data and normals: Asheville, North Carolina, U.S. De-partment of Commerce, National Oceanic and Atmospheric Administration, National Environmental Satellite Data and Information Service, National Climatic Data Center. http://www7.ncdc.noaa.gov/CDO/cdo (accessed 12 Dec. 2012)

Newman, B., A. Tanweer, and T. Kurttas. 2008. IAEA standard operating proce-dure for the DT-100 liquid-water isotope analyser. IAEA, Vienna.

Nutter, W.L. 1964. Determination of the head-discharge relationship for a sharp-crested compound weir and a sharp-crested parabolic weir. Ph.D. diss., Pennsylvania State University, University Park.

O’Driscoll, M.A., D.R. DeWalle, K.J. McGuire, and W.J. Gburek. 2005. Seasonal 18O variations and groundwater recharge for three landscape types in

central Pennsylvania, USA. J. Hydrol. 303(1–4):108–124. doi:10.1016/j.jhy-drol.2004.08.020

Pearce, A., M. Stewart, and M. Sklash. 1986. Storm runoff generation in humid headwater catchments 1. Where does the water come from? Water Resour. Res. 22(8):1263–1272. doi:10.1029/WR022i008p01263

Reich, E.G. 1966. A rain study of a small drainage basin in central Pennsylvania. M.S. diss., Pennsylvania State University, University Park.

Rice, K.C., and G.M. Hornberger. 1998. Comparison of hydrochemical tracers to estimate source contributions to peak flow in a small, forested, headwater catchment. Water Resour. Res. 34(7):1755–1766. doi:10.1029/98WR00917

Shapiro, S.S., and M.B. Wilk. 1965. An analysis of variance test for normality (complete samples). Biometrika 52(3/4):591–611. doi:10.2307/2333709

Sklash, M.G., and R.N. Farvolden. 1979. The role of groundwater in storm run-off. J. Hydrol. 43:45–65. doi:10.1016/0022-1694(79)90164-1

Sklash, M.G., R.N. Farvolden, and P. Fritz. 1976. A conceptual model of wa-tershed response to rainfall, developed through the use of oxygen-18 as a natural tracer. Can. J. Earth Sci. 13:271–283. doi:10.1139/e76-029

Tang, K., and X. Feng. 2001. The effect of soil hydrology on the oxygen and hy-drogen isotopic compositions of plants’ source water. Earth Planet. Sci. Lett. 185:355–367. doi:10.1016/S0012-821X(00)00385-X

Vanclooster, M., M. Javaux, and J. Vanderborght. 2005. Solute transport in soil at the core and field scale. In M.G. Anderson and J.J. McDonnell, editors, En-cyclopedia of hydrological sciences. John Wiley & Sons, New York. p. 1–13.

Vogel, T., M. Sanda, J. Dusek, M. Dohnal, and J. Votrubova. 2010. Using oxy-gen-18 to study the role of preferential flow in the formation of hillslope runoff. Vadose Zone J. 9(2):252–259. doi:10.2136/vzj2009.0066

Whillans, I., and P. Grootes. 1985. Isotopic diffusion in cold snow and firn. J. Geophys. Res. 90(D2):3910–3918. doi:10.1029/JD090iD02p03910

White, A.F., M.S. Schulz, D.V. Vivit, A.E. Blum, D.a. Stonestrom, and J.W. Hard-en. 2005. Chemical weathering rates of a soil chronosequence on granitic alluvium: III. Hydrochemical evolution and contemporary solute fluxes and rates. Geochim. Cosmochim. Acta 69(8):1975–1996. doi:10.1016/j.gca.2004.10.003

Winograd, I.J., A.C. Riggs, and T.B. Coplen. 1998. The relative contribu-tions of summer and cool-season precipitation to groundwater recharge, Spring Mountains, Nevada, USA. Hydrogeol. J. 6(1):77–93. doi:10.1007/s100400050135

Zimmermann, U., D. Ehhalt, and K. Münnich. 1968. Soil-water movement and evapotranspiration: Changes in the isotopic composition of the water.. In IAEA, editors, Isotopes in hydrology. International Atomic Energy Agency. Vienna. p. 567–584.

Documents

Special Section: Frontiers Spatiotemporal Patterns of ...plsullivan/thomasetal_2013.pdf · along with other chemical tracers to conceptualize catchment hydrology (Holmes, 2011; Jin