Effects of soil temperature and depth to ground water on first-year growth of a dryland riparian phreatophyte, Glycyrrhiza lepidota (American licorice).
Climatic warming has the potential to directly affect all life stages of perennial riparian plants, starting with temperature-dependent timing of germination of seeds (Baskin and Baskin, 1998) and the pattern of growth of seedlings (Nord and Lynch, 2009). Indirect effects from climatic warming will accrue through change in the local flow regime, which often controls the dynamics of ground water in dryland riverine ecosystems. These indirect effects can thus be identical to the direct hydrological effects of some forms of management of dryland stream flow. Alteration of flow, regardless of source, can in turn influence temperatures of floodplain soil through its influence on patterns of hyporheic flow and associated advective heat-exchange (Gerecht et al., 2011), as well as moisture content of riparian soil (Brady and Weil, 2002; Geiger et al., 2003). In North America, our speciesspecific knowledge of response of dryland riparian plants to soil temperature and ground-water hydrology is very limited for all but a few ecologically dominant woody species. This lack of information hampers our understanding of the full effects of management of water resources and our ability to make specific predictions about effects of climatic change on vegetation (Perry et al., 2012).
We present the first climate-related autecological study for the facultatively riparian plant, Glycyrrhiza lepidota (Fabaceae; American licorice), a nitrogen (N)-fixing [C.sub.3] herb widespread in the United States, southern Canada, and northern Mexico (United States Department of Agriculture, plants database, http://plants.usda.gov). Glycyrrhiza lepidota is the only North American member of the globally distributed genus and is an important source of nectar and food for native butterflies and other insects. The N-fixing capability of the species likely adds to its importance along dryland river floodplains of western North America, which may commonly be N-limited (Holmes et al., 1994; Adair and Binkley, 2002; but see Andersen et al., 2014). Glycyrrhiza lepidota can develop a long taproot, which provides scour resistance and allows the plant to access a relatively deep water table. Thus, the species may serve as a model organism for herbaceous riparian phreatophytes. Although published demographic studies are rare, first-year growth appears to be entirely devoted to nonreproductive structures, with flowers first appearing during the second growing season (Platt, 1988, in Bender et al., 2000). Reproduction is sexual and vegetative through rhizomes and adventitious shoots from roots.
We used mesocosms incorporating subirrigation to examine the effects of temperature of soil (two levels) and position of the water table (two levels) on growth of G. lepidota through its first growing season. We hypothesized that relatively moist and warm soil would be most favorable and predicted fastest growth of shoots and roots in such a soil, unless our temperatures exceeded those normally encountered in nature or volume of soil became constraining. Further, because formation of nodules and N-fixation also might be promoted in the relatively warm and moist soil (Zahran, 1999) and rate of N-fixation of a plant is proportional to the biomass of its roots (Holter, 1978), we hypothesized a lower need and uptake of soil N by plants in the moist and warm treatment. Thus, we predicted that inorganic soil-N levels at the end of the growing season would be highest in the moist and warm treatment.
Although many studies have examined growth of dryland riparian plants under controlled levels of ground water (e.g., Kranjcec et al., 1998; Amlin and Rood, 2002; Sher and Marshall, 2003), few controlled the level of dissolved oxygen (DO) in ground water. Dryland alluvial ground water is commonly anoxic (D. C. Andersen, pers. observ.; Vinson et al., 2007) because of the rapid microbial uptake of oxygen from its predominant source water, infiltrating (hyporheic) surface water. Because many riparian plants have roots that function poorly under anoxic conditions and that can be killed by prolonged exposure to anoxia (Naumburg et al., 2005), we maintained low DO in our mesocosm water to more accurately mimic conditions of a natural dryland floodplain.
MATERIALS AND METHODS--We mimicked conditions of floodplain soil and ground water using 122-cm tall soil-filled polyvinyl chloride (PVC) tubes (=pots) set vertically into modified polyethylene 208-L barrels (BayTec Containers, Model FR401 with installed bulk-head fitting) partially filled with anoxic water. We controlled the depth from the surface of the soil to the water table ([D.sub.W]) in each pot at either ca. 100 or ca. 63 cm using PVC standpipes attached to the bulk-head fitting. The mesocosms were built inside a greenhouse near Denver, Colorado, that featured translucent fiberglass roof and side panels and in which air temperature (but not humidity) was quasi-controlled by a heating unit and ventilation fan. Ambient humidity in the greenhouse, although not measured, was likely consistently low, reflecting the semiarid climate of Denver.
We constructed the mesocosms by placing seven empty pots, constructed from 10.2-cm inside-diameter, thin-walled sewer pipe, into each barrel in a vertical position (Fig. 1). Twelve barrels were placed next to one another along the centerline of the SE-NW trending greenhouse, and the pots (n = 84) were systematically filled to ca. 1 cm from their top with screened topsoil. Once pots were filled with soil, no pot or barrel was moved. One or another of the four combinations of treatments was systematically assigned to each barrel (Fig. 1).
Soil in pots was characterized as a slightly alkaline, slightly saline, sandy clay loam with low organic matter (0.7-0.9%, n = 3) and nutrient levels (10 ppm N[O.sub.3]-N, ca. 9 ppm available phosphorus, and ca. 85 ppm available potassium; Colorado State University Soil, Water and Plant Testing Laboratory, Fort Collins, Colorado). Final bulk density was estimated to be 1.55 air-dry g/ [cm.sup.3] and moisture content was 3.8%. We estimated the height of the capillary fringe (CF) in the soil using a 90-cm long, clear plastic cylinder (4.8-cm inside diameter) filled with dry soil in the same manner as pots. We set the bottom of the vertical tube into a reservoir of water maintained at a constant depth and measured the position of the top of the CF (junction of light and dark soil) relative to the surface of the reservoir after [greater than or equal to] 4 days.
During the process of filling pots with soil, we installed a gypsum soil-moisture block (Delmhorst Instrument Company, Towaco, New Jersey, model GB-1) in one randomly selected pot in each barrel; two barrels had a second block placed in a different, randomly selected pot. We also installed a temperature sensor and cable (Campbell Scientific, Logan, Utah, model 107) in one randomly selected pot in each of seven systematically selected barrels. No pot contained a sensor and a moisture block. Gypsum blocks were installed such that the bottom was 60 cm below the rim of the pot, and temperature sensors (installed vertically) had their tip at 58 cm below the rim of the pot. We collected data on moisture of soil manually ([greater than or equal to] 1/week, Delmhorst Model KS-D1 moisture meter), whereas we collected data on temperature of the soil using a datalogger (Campbell Scientific, model CR800) sampling at 5-min intervals and averaging over each 3-h interval and daily. The datalogger restricted simultaneous monitoring of the temperature of soil to four of the seven instrumented pots.
All water added to barrels was from the municipal supply and typically cool to cold. We first added water to the barrels on 28 January 2011, with all barrels fitted with a 20-cm standpipe ([D.sub.W] = 100 cm or large) to have uniform conditions for the germination of seeds. A 57-cm standpipe, reducing [D.sub.W] to 63 cm ([D.sub.W] = small), was installed on half the barrels on 8 March 2011, when seedlings were present in all but four pots.
We used submersible, aquaria heating-elements (TransWorld Aquatic Enterprises, Inglewood, California, TrueTemp[R] Titanium T3-300) to heat water in half of the barrels (Fig. 1), with the expectation that the heated water would then warm the soil in the pots in those barrels to a temperature ([T.sub.S] = warm) above that in the unheated barrels ([T.sub.S] = cold). Heating began on 3 February 2011 (all barrels with 20-cm standpipes), with thermostats set to 15.6[degrees]C (60[degrees]F; accuracy [+ or -] 0.5[degrees]F). Thermostats were raised to 18[degrees]C (65[degrees]F) on 15 February and then 21[degrees]C (70[degrees]F) on 18 February, where they remained for the duration of the experiment. We monitored water temperature ([T.sub.W]) in each barrel at 30-min intervals using Water Temp Pro v2 dataloggers (Onset Computer Corporation, Bourne, Massachusetts, HOBO model U22-001; Fig. 1). The four combinations of treatments were: [T.sub.S] = cold and [D.sub.W] = large (CL); [T.sub.S] = cold and [D.sub.W] = small (CS); [T.sub.S] = warm and [D.sub.W] = large (WL); and [T.sub.S] = warm and [D.sub.W] = small (WS).
We added a somewhat recalcitrant carbon source with little or no nutrient value (cornstarch) to the barrel water (5 g/L) on 28 March to stimulate microbial production and thereby a high biological oxygen demand (Goetzman, 1990). A second dose (5 g/L) was added 16 August 2011. We monitored levels of DO sporadically using a handheld DO meter (YSI, Yellow Springs, Ohio, model YSI 85/10 DO meter.
Only natural light entered the greenhouse. We continuously monitored levels of photosynthetically active radiation (PAR) near the barrels using the datalogger and a single sensor (LI-COR Quantum Sensor, LI-190SB). Air temperature of the greenhouse was monitored using HOBO dataloggers attached to the outside, shaded (northeast) portion of each of barrels 1 and 11 (Fig. 1).
Our intent was to grow one plant from seed in each of the 84 growing pots. We planted inoculated (but unscarified) seed on 14January 2011. We also planted seeds at this time, and up to 6 weeks later, in 5-cm deep trays containing the same soil as that in pots. We transplanted seedlings from these trays into pots in which no germinants appeared within 2-3 weeks of sowing or resowing. The last reseeding of pots occurred 31 January, and the last transplantation occurred on 18 April. Transplantation always involved seedlings at the cotyledon stage (height of 1-2 cm).
We top-irrigated pots with small amounts of water (usually 35 or 70 ml) every 1-4 days as judged needed until mid-March, when the largest seedlings were ca. 3-cm tall and the second set of true leaves was developing on most plants. Each pot also was fertilized (Miracle-Gro All Purpose Plant Food) via top-watering on three occasions (50 ml/occurrence) in late March and early April, with total fertilization reaching the equivalent of 21, 11, and 16 g/[m.sup.2] for nitrogen, phosphorus, and potassium, respectively. We applied a liquid insecticide (Safer End All) on 23 March 2011, after discovering webbing, possibly from spider mites, on some seedlings. A second spraying occurred on 25 March.
Height of seedlings (H) was measured weekly beginning 17 March. When multiple seedlings were present in a pot, all smaller seedlings were clipped at ground level if the largest measured [greater than or equal to] 15 cm in height. Regrowth of clipped individuals was rare.
We terminated the study on 9 September 2011 by draining barrels and disconnecting heaters. We harvested shoots on 15 September, and, on 26-27 September, we cut each PVC pipe in half longitudinally and sampled soil and excavated root systems. We collected soil (100 ml) from a depth of 50 cm in each pot containing a plant and pooled samples from each barrel, resulting in 12 samples for analysis. We also collected samples of soil at the same depth from one or two empty pots (i.e., pots that contained no plant at the end of the study) in each combination of treatments (one pot per barrel) and analyzed them individually (total n = 6). These soils served as controls for comparison with levels of nitrogen in pots with plants. We carefully removed the exposed root system and removed adhering soil by dipping the root into a water bath. We dried (60[degrees]C) shoots and roots (including rhizomes) to constant weight to determine biomass.
We searched for any systematic environmental gradient that might be affecting growth along the linear arrangement of barrels by comparing heights at a point in time ([H.sub.T]) among the three barrels within each combination of treatments, using one-way analysis of variance (ANOVA) with position of barrel as the independent variable. Because a preliminary analysis indicated the family of height-growth-curves did not meet the sphericity requirement for repeated-measures ANOVA (Greenhouse-Geisser and Huynh-Feldt epsilon values = ca. 0.06), we used an alternative statistical test to test the growth curves for treatment effects (Elso et al., 2004; http://bioinf.wehi.edu.au/software/ compareCurves/). The test statistic (mean t) was the two-sample t-statistic comparing [H.sub.T] in two groups at each weekly measurement, averaged over the 25-week-long experiment. We first made pairwise comparisons among the three replicate barrels in each treatment, considering each pot as an independent experimental unit. We then pooled pots (i.e., individual plants) if no significant difference was detected among the barrels. Next, we made pairwise comparisons among the four kinds of treatments. We generated P-values for the test statistic using [greater than or equal to] 4,000 permutations of the dataset.
We also tested for treatment effects on HT on three dates using two-factor ANOVA, incorporating a covariate if deemed appropriate from the prior analysis. We performed similar analyses for final shoot biomass (BS), root biomass ([B.sub.R]), total biomass (B = [B.sub.S] + [B.sub.R]), and root:shoot ratio (R = [B.sub.R]/[B.sub.S]). We tested for differences in soil nitrogen (each of N[O.sub.3]-N, NH4-N, and total N) among treatments at the end of the study using twofactor ANOVA. We also compared soil nitrogen in pots with and without plants using a two-sample t-test. With the exception of analysis of the growth curves, all statistical analyses were conducted using SYSTAT[R] 11 (Systat Software Inc., San Jose, California). Means [+ or -] 1 SE are presented, with sample size in parentheses unless clear from the text. We assumed statistical significance at [alpha] = 0.05.
RESULTS--Four independent trials indicated the height of the capillary fringe in the pots of soil likely reached the surface of the soil in the treatments [D.sub.W] = S but not in the treatments [D.sub.W] = L. Two trials indicated CF [greater than or equal to] 45 cm and [greater than or equal to] 57 cm after 4 and 7 days, respectively. A third trial indicated CF [greater than or equal to] 61 cm after ca. 1 week. A single multi-week trial indicated CF = 77 cm after 4 weeks. The readings from the KS-D1 moisture block meter increased to [greater than or equal to] 96 (ca. 85% saturation) within 4 days following the first addition of water to the barrels in late January, when all barrels had water at a level 40 cm below the position of the block (i.e., [D.sub.W] = L in all pots). Subsequent readings remained [greater than or equal to] 96 in 11 of the 12 barrels. The exception was a CL barrel that happened to have two pots with moisture blocks. The readings for one pot declined to 83 in early July and reached 78 by the end of the study. The reading for the second pot began decreasing in mid-August and was 90 at the end of the study.
The data on soil temperature indicated that our warming method was effective. The difference in mean [T.sub.S] in pots between heated and unheated barrels was maximal in early spring (2.78C) and decreased to a minimum (0.5[degrees]C) in early July before increasing to 1.4[degrees]C in mid-August (Fig. 2). Earlier (12-17 March) monitoring of [T.sub.S] in two other barrels with different combinations of treatments (CL and WS) also showed higher [T.sub.S] in the heated barrel (data not shown). These data and the similarities in [T.sub.W] within [T.sub.S] treatments (data not shown) provide confidence that [T.sub.S] in heated barrels was raised relative to the unheated controls in [D.sub.W] treatments. Dissolved oxygen of water in barrels decreased to < 1 mg/L within weeks after addition of the cornstarch and remained at low levels throughout the study (Fig. 2).
The maximum instantaneous level of PAR in the greenhouse was 1,038 [micro]mol photons/s/[m.sup.2], recorded on 16 July, whereas the maximum 3-h average PAR was 713 [micro]mol photons/s/[m.sup.2] (8 August). Total daily PAR during the months of June, July, and August ranged from 7.1-18.2 mol photons/[m.sup.2]/d; the mean was 13.55 [+ or -] 0.32 mol photons/[m.sup.2]/d.
Seedlings first appeared in pots about 10 days after sowing, but rate of germination and early survival were poor. Fourteen of 84 pots (17%) were reseeded, and 67% of the pots eventually received at least one transplant. Sixty-three of the 84 pots contained a live plant at the end of the experiment, but one plant in each of barrels 10 (CS treatment) and 11 (WL) showed extremely poor vigor throughout the experiment and were deleted from the analyses. Of the 61 plants used in the analyses, 34 (56%) were transplants. Each barrel contributed three or more plants to the analyses.
Seedlings established in pots from seed and those transplanted into pots grew at comparable rates. Eleven of the 12 barrels contained at least one plant established from seed. Within-treatment t-tests (two-sample with pooled variance) comparing mean total biomass ([B.sub.T]) of transplants to that of plants established from seed revealed a difference only in the CS treatment (P = 0.034). Inspection of the data from the CS treatment indicated the difference stemmed largely from one unusually large transplant and one unusually small plant established from seed, both growing in barrel 6. Given these results, we consider it reasonable to ignore origin of plants in all analyses of treatment effects on growth.
Rate of growth in height of shoots was slow until June and then dramatically increased (Fig. 3a). Pairwise comparisons of growth curves among the three barrels within a combination of treatments suggested patterns of growth were equivalent in each of the four assessments (P [greater than or equal to] 0.33); thus, barrels were pooled for pairwise comparisons among the combinations of treatments. That analysis indicated a highly significant effect from [D.sub.W] regardless of [T.sub.S], whereas no effect from [T.sub.S] was detected in either comparison involving constant [D.sub.W] (Table 1). Corroborating this analysis, a two-way ANOVA of height of plants on 3 June (n = 68) and 8 July (n = 65) indicated a [D.sub.W] effect (P < 0.001) but no [T.sub.S] or interaction ([D.sub.W] x [T.sub.S]) effect. Plants in the pots with [D.sub.W] = S were more than twice as tall as those in the pots with [D.sub.W] = L by 8 July (Fig. 3b). This advantage in height declined over late summer; however, the same ANOVA on final (9 September) heights (n = 61) indicated no significant difference among the treatments, although the difference between [D.sub.W] treatments was nearly significant (P = 0.058). The effect of [D.sub.W] on height became significant (P = 0.027) when position of barrels was used as a covariate to account for a spatial gradient most apparent in biomass of shoots. The covariate was significant (P = 0.020).
The spatial gradient in response of plants was evident in shoot biomass ([B.sub.S]), with means for combinations of treatments consistently lowest in the highest numbered (NW-most) barrel (Fig. 4), i.e., the barrel closest to the fan circulating air in the greenhouse. This pattern also was evident in [B.sub.R] for three of the four combinations of treatments (Fig. 4a). The gradient of response could have been due to variation in air currents, undetected variation in intensity of light or the modest ([less than or equal to] 0.34[degrees]C) differences in mean seasonal air temperature measured along the array of barrels. One-way ANOVA testing for a position effect in [B.sub.S] among the three barrels in each combination of treatments indicated a significant difference only in the CL combination (P = 0.024). Nevertheless, because of the uniformity of the trend among treatments, we assume the spatial gradient was real and affected all combinations. We, therefore, tested for the presence of a significant interaction between the covariate (barrel position) and each of [D.sub.W] and [T.sub.S] (i.e., assumption of homogeneity of slopes in analysis of covariance, ANCOVA), and finding none, conducted two-way ANOVA with barrel position as a covariate. The results indicated a [D.sub.W] effect on [B.sub.S] (P = 0.052), but again no effect from [T.sub.S] was detected (Table 2).
With one exception, roots ended in a narrow zone a few centimeters above the water table. The exception was an unusual plant in CS having a relatively small shoot and the largest root:shoot ratio (R = 3.50) observed in the study. Structure of roots differed among the [D.sub.W] treatments, with branches off the main taproot tending to be relatively sparse when [D.sub.W] = L, whereas they formed a dense mat above the water table when [D.sub.W] = S. Mean [B.sub.R] was consistently higher than mean [B.sub.S] (Fig. 4a). Two-way ANCOVA examining effects of [D.sub.W]and [T.sub.S] on [B.sub.R] produced results paralleling those for BS, as did the same analysis on total biomass (Table 2) and root:shoot ratio. Root:shoot ratio was higher for pots with [D.sub.W] = S than for pots with [D.sub.W] = L (2.19 versus 1.85). Nearly half (42%) of the plants in pots with [D.sub.W] = S had at least one rhizome present (included in the measurement of biomass of the root), whereas no plant in the treatments with [D.sub.W] = L produced a rhizome. Nodules were concentrated on root branches, with few on the taproot.
A hardpan ranging from 6-40 cm in thickness and extending to the surface of the soil developed in 12 of the pots with [D.sub.W] = S. The remaining pots with plants, including all pots with [D.sub.W] = L, featured a friable surface soil at the end of the study. The thickness of hardpan was positively related to biomass of shoots (Fig. 5; [r.sup.2] = 0.56, P = 0.005) as well as biomass of roots (P = 0.022) and total biomass (P = 0.015).
We found a significant effect from [D.sub.W] on final level of soil nitrate in pots containing Glycyrrhiza, but no [T.sub.S] effect or interaction ([D.sub.W], P < 0.001; [T.sub.S], P = 0.098; [D.sub.W] x [T.sub.S]: P = 0.69). Mean N[O.sub.3]-N in the treatments with [D.sub.W] = S was more than double the level in the treatment [D.sub.W] = L (Table 3; Fig. 4b). No effect on either soil ammonium or total N was detected. Mean N[O.sub.3]-N also varied with [D.sub.W] in the pots without plants but in a pattern opposite to that found in pots with plants (Table 3). Whereas nitrate levels in pots with and without plants were equal in the treatment [D.sub.W] = S, there was a five-fold difference in the treatment [D.sub.W] = L, with empty pots having the higher concentration of nitrate (Table 3). Total N in empty pots was about twice the level present in pots with plants (treatments pooled, t-test with separate variance, P = 0.032), whereas levels of total carbon did not differ (P = 0.65). As a result, the C:N ratio in empty pots was about half that found in pots with plants (P = 0.002).
DISCUSSION--The comparisons growth curves (height) and final analyses of biomass indicated that depth to ground water, but not soil temperature, affected growth of Glycyrrhiza lepidota over its first growing season. Three factors linked to [D.sub.W] stand out as possible causes of the 19 and 31% lower growth of shoots and roots, respectively, in the [D.sub.W] = L treatment. First, the uppermost soil in pots with [D.sub.W] = L would have become relatively dry after we ended top-irrigation, because this soil was above the capillary fringe and evaporative rate in the greenhouse, reflecting the regional semiarid climate, was consistently high. This dry soil zone must have been restricted to the upper 30 cm or so of the soil in the pot, however, based on the height of the capillary fringe and the uniformly high readings from the moisture block. Growth of seedlings in the treatment [D.sub.W] = L may have been constrained by low soil moisture until roots grew through this zone, retarding the early summer onset of rapid accumulation of biomass (Fig. 3a). The data on height indicates that the effect of [D.sub.W] on growth was in fact most pronounced early in the growing season.
Secondly, growth of plants in [D.sub.W] = L may have been constrained by low availability of N because of low symbiotic microbial activity. Deficiency in soil moisture strongly reduces [N.sub.2] fixation because initiation, growth, and activity of nodules are more sensitive to water stress than is metabolism of roots and shoots (Serraj et al., 1999; Zahran, 1999). Formation of nodules and rates of growth would have had a strong influence on availability of N to the plant given low availability of inorganic soil N. In fact, the final values of total soil N in pots without plants (Table 3) remained at the low end for dryland riparian soils (despite our fertilization), comparable to values Adair et al. (2004) reported for the earliest stage of floodplain succession in semiarid northwestern Colorado (0.06%).
Finally, the growing taproots would have required more time to reach the water table in the treatment [D.sub.W] = L, even if rates of growth of taproots were equal in the two treatments of [D.sub.W]. This could have delayed accumulation of biomass if proliferation of branch roots was triggered by contact of roots with the water table. Support for this idea is provided by Imada et al. (2008), who found that growth of fine roots in phreatophytic Populus alba was stimulated in the layer of soil just above a stationary water table. A simultaneous delay in formation of nodules may have occurred, because nodules appeared to be concentrated on the fine branch roots, with few on the taproot. Regardless of the mechanism involved, the effect of [D.sub.W] suggests that growth of seedlings of G. lepidota is highly sensitive to low soil moisture.
Although we detected no effect from [T.sub.S], the North American distribution of G. lepidota suggests that the species is in fact sensitive to environmental temperature. Whereas G. lepidota is found in uplands in mesic regions, it is strictly riparian in drylands. In Arizona and New Mexico, for example, G. lepidota occurs in montane wetlands (Sivinski, 2007) but is rare or absent at the warmest mid-elevation and low-elevation riparian sites (Wolden et al., 1995; Beauchamp and Shafroth, 2011; also see United States Department of Agriculture Forest Service Fire Effects Information System: http://www.fs. fed.us/database/feis/plants/forb/glylep/all.html).
Our failure to find an effect from [T.sub.S] may be partially a result of our choice of 21[degrees]C for the set point of the immersion heater. Minimum [T.sub.W] in unheated barrels reached this threshold in June, with the result that heaters were idle during the warmest part of the growing season. A higher set point, however, could have increased [T.sub.S] of heated barrels to levels rarely, if ever, observed in a natural riparian setting. The mean soil temperatures in the heated and unheated mesocosms ([T.sub.S] = 27-31[degrees]C; Fig. 2) were higher than maximum temperatures observed at a shallower depth (40 cm) on the floodplain of the free-flowing Yampa River in semiarid northwestern Colorado (16-18[degrees]C; D. C. Andersen, pers. observ.). Soil temperature at a 60-cm depth on that floodplain would be even cooler. However, our pot [T.sub.S] values were similar to maximums measured at a 40-cm depth on the floodplain of the regulated Green River, also in northwest Colorado (25-32[degrees]C; D. C. Andersen, pers. observ.). Dryland riparian soil temperatures are poorly documented and effects on the soil thermal regime from hydrologic alteration remain largely unexplored.
Our finding of a strong effect of [D.sub.W] supports the contention of Perry et al. (2012) that the indirect effects of climatic change in dryland riparian ecosystems may be as strong or stronger than direct effects. Any alteration in depth to ground water will be a response to stream flow, which on snowmelt-dominated alluvial rivers is determined primarily by climatic conditions in distant mesic headwater areas. Thus, the change in depth to ground water is an indirect effect, mediated by streamflow. Interactions and feedbacks, however, are likely. For example, changes in [D.sub.W] could affect vegetative structure and thereby ground surface shading, which would affect [T.sub.S].
Our data support the concept that roots of G. lepidota will not grow into anoxic soil. However, the presence of mature plants on floodplains suggests that existing roots tolerate inundation in anoxic ground water to some degree. Whereas riparian seedlings are unlikely to survive where surface or near-surface soil is consistently saturated, established plants can clearly tolerate sites temporarily saturated during floods. Seeds of G. lepidota ripen in autumn and presumably germinate the following spring. In dryland riparian environments, a spring or summer flood pulse that either triggers or coincides with germination would promote establishment if it resulted in a moist surface soil to support germinant growth but the flood recession and accompanying soil drying was not so rapid that it outpaced growth of roots of seedlings to the deeper, permanent source of moisture needed for the post-flood growing season. This set of physical constraints is analogous to the well-documented set of requirements for establishment of western riparian cottonwood (Populus; Mahoney and Rood, 1998; Cooper et al., 1999). The fact that formation of rhizomes was evident only in the pots with [D.sub.W] = S suggests that G. lepidota can rapidly initiate vegetative expansion when soil conditions are favorable. Thus, the most favorable dryland riparian site may be one featuring a moderately shallow water table, few constraints on lateral growth, and a soil texture that drains at a moderate pace.
We attribute the hardpan that developed in some of the pots with [D.sub.W] = S to proliferation of foliage that increased rates of water-uptake to levels sufficient to capture much or all of the ground water being moved upward by capillary action. This process would desiccate and thereby harden the soil above the roots involved (Passioura, 2002). An ability to desiccate surface soils above its root system, if present in natural dryland riparian settings, implies that G. lepidota may be able to restrict other plants from establishing or growing under or near it. Few studies have examined the effect from dryland phreatophytes on moisture level in the soil above their root systems. Schade and Hobbie (2005) found no effect from phreatophytic velvet mesquite trees (Prosopis velutina) on surface (0-10 cm) soil moisture in a riparian zone in the Sonoran Desert, but the depth of the water table relative to height of the capillary fringe was not noted. Busch and Smith (1993) found no difference in levels of surface soil moisture in a comparison of burned and unburned riparian areas but noted that phreatophytes would be unlikely to substantially deplete water in surface soils where average depth of the water table is deep (i.e., where [D.sub.W] > CF). Indirect evidence of a phreatophyte intercepting rising capillary water and thereby affecting higher soils comes from Kushiev (2005), who reported that plantings of the Eurasian G. glabra reduced the water table and reduced salt content of soils in a salinized portion of the Aral Sea Basin. Such a reduction in surface soil salts would result from interception of saline capillary water coupled with leaching of surface soil by infiltrating precipitation.
We found the highest root:shoot ratio in the treatment [D.sub.W] = S. In contrast, Shafroth et al. (1995) found that root:shoot ratio tended to be positively related to [D.sub.W] after the first growing season in plains cottonwood, P deltoides Marshall subsp. monilifera (Aiton). This discrepancy may be related to how the architecture of roots responds to soil moisture. In the case of G. lepidota, exposure of taproots to anoxic, saturated soil, which occurred earliest in pots with [D.sub.W] = S, may have stimulated branching of roots and expansion of fine roots. It also may have initiated production of rhizomes, which were noted solely in the treatment [D.sub.W] = S.
Our prediction that the level of inorganic soil N would be highest in the moist, warm treatment was partially confirmed, in that the level of N[O.sub.3]-N was greatest in the pots with [D.sub.W] = S (Table 3; Fig. 4b) and soil N[H.sub.4]-N showed a trend in the predicted direction (Table 3). We found no effect from [T.sub.S]. The similarity in values of total N (Table 3) suggests a relatively high concentration of organic-N in soils with [D.sub.W] = L. A possible explanation is that rate of mineralization was higher in the in the treatment [D.sub.W] = S, which would have reduced organic N and contributed to the larger pool of inorganic N in that treatment.
The difference in level of nitrate between pots with and without plants in the treatment [D.sub.W] = L (Table 3) suggests, not surprisingly, that the taproot depletes soil mineral N as it grows downward. This N pool would eventually be replaced and enhanced through leaf-fall and subsequent decomposition, although a time lag measured in years could be involved in a dryland riparian environment (Andersen and Nelson, 2006). The difference in levels of nitrate in the empty pots at different [D.sub.W] (Table 3) may be due to differences in microbial activity, e.g., greater assimilation of nitrate by bacteria and fungi in the moister pots of treatment [D.sub.W] = S.
Our results indicate that seedlings of G. lepidota in dryland riparian settings grow fastest in unsaturated soil but where water is accessible from the capillary fringe (i.e., over a moderately shallow water table). The results also suggest a potential for the species, once phreatophytic, to affect competing vegetation by intercepting rising capillary water and thereby desiccating soils above the root system. Although we found no effect from soil warming within the temperature range we examined, further experiments manipulating air and soil temperatures may demonstrate that the distribution of G. lepidota in dryland riparian sites is restricted by high summer temperatures. Finally, our results suggest that nodulation is at least retarded when seedlings of G. lepidota must grow through a relatively dry soil to access a deep water table. It remains to be seen whether this would significantly reduce the potential of G. lepidota to improve levels of plant-available soil N in N-limited dryland riparian environments. Because climatic change as well as regulation of stream flow may dry surface soils and cause declines in floodplain water tables, each has the potential to adversely affect streamside G. lepidota and alter associated communities of plants. Our results have implications for managers of water resources interested in developing environmental flows to promote or sustain populations of G. lepidota, either in riparian areas where the species now occurs or as part of efforts to restore vegetation in degraded areas.
This work was funded by the Science and Technology Program of the Bureau of Reclamation and the United States Geological Survey. We thank the anonymous reviewers for helpful comments on previous versions of the manuscript. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the United States Government.
ADAIR, E. C., AND D. BINKLEY. 2002. Co-limitation of first year Fremont cottonwood seedlings by nitrogen and water. Wetlands 22:425-429.
ADAIR, E. C., D. BINKLEY, AND D. C. ANDERSEN. 2004. Patterns of nitrogen accumulation and cycling in riparian floodplain ecosystems along the Green and Yampa rivers. Oecologia 139:108-116.
AMLIN, N. M., AND S. B. ROOD. 2002. Comparative tolerances of riparian willows and cottonwoods to water-table decline. Wetlands 22:338-346.
ANDERSEN, D. C., AND S. M. NELSON. 2006. Flood pattern and weather determine Populus leaf litter breakdown and nitrogen dynamics on a cold desert floodplain. Journal of Arid Environments 64:626-650.
ANDERSEN, D. C., D. J. COOPER, AND K. NORTHCOTT. 2007. Dams, floodplain land use, and riparian forest conservation in the semiarid Upper Colorado River Basin, USA. Environmental Management 40:453-475.
ANDERSEN, D. C., E. C. ADAIR, S. M. NELSON, AND D. BINKLEY. 2014. Can nitrogen fertilization aid restoration of mature tree productivity in degraded dryland riverine ecosystems? Restoration Ecology 22:582-589.
BASKIN, C. C., AND J. M. BASKIN. 1998. Seeds: ecology, biogeography, and evolution of dormancy and germination. Academic Press, New York.
BEAUCHAMP, V. B., AND P. B. SHAFROTH. 2011. Floristic composition, beta diversity, and nestedness of reference sites for restoration of xeroriparian areas. Ecological Applications 21:465-476.
BENDER, M. H., J. M. BASKIN, AND C. C. BASKIN. 2000. Age of maturity and life span in herbaceous, polycarpic perennials. Botanical Review 66:311-349.
BRADY, N. C., AND R. R. WEIL. 2002. The nature and properties of soils. Thirteenth edition. Prentice Hall, Upper Saddle River, New Jersey.
BUSCH, D. E., AND S. D. SMITH. 1993. Effects of fire on water and salinity relations of riparian woody taxa. Oecologia 94:186-194.
COOPER, D. J., D. M. MERRITT, D. C. ANDERSEN, AND R. A. CHIMNER. 1999. Factors controlling the establishment of Fremont cottonwood seedlings on the upper Green River, USA. Regulated Rivers: Research & Management 15:419-440.
ELSO, C. M., L. J. ROBERTS, G. K. SMYTH, R. J. THOMSON, T. M. BALDWIN, S. J. FOOTE, AND E. HANDMAN. 2004. Leishmaniasis host response loci (lmr1-3) modify disease severity through a Th1/Th2-independent pathway. Genes and Immunity 5:93-100.
GEIGER, R., R. H. ARON, AND P. TODHUNTER. 2003. The climate near the ground. Sixth edition. Littlefield Publishers, Rowman and Lanham, Maryland.
GERECHT, K. E., M. B. CARDENAS, A. J. GUSWA, A. H. SAWYER, J. D. NOWINSKI, AND T. E. SWANSON. 2011. Dynamics of hyporheic flow and heat transport across a bed-to-bank continuum in a large regulated river. Water Resources Research 47:W03524.
GOETZMAN, J. S. 1990. The impact of internal phosphorus loading on the restoration of Trout Lake. Environmental Progress 9:69-72.
HOLMES, R. M., S. G. FISHER, AND N. B. GRIMM. 1994. Parafluvial nitrogen dynamics in a desert stream ecosystem. Journal of the North American Benthological Society 13:468-478.
HOLTER, V. 1978. Nitrogen fixation of four legumes in relation to above-ground biomass, root biomass, nodule number, and water content of the soil. Oikos 31:230-235.
IMADA, S., N. YAMANAKA, AND S. TAMAI. 2008. Water table depth affects Populus alba fine root growth and whole plant biomass. Functional Ecology 22:1018-1026.
KRANJCEC, J., J. M. MAHONEY, AND S. B. ROOD. 1998. The responses of three riparian cottonwood species to water table decline. Forest Ecology and Management 110:77-87.
KUSHIEV, H., A. D. NOBLE, I. ABDULLAEV, AND U. TOSHBEKOV. 2005. Remediation of abandoned saline soils using Glycyrrhiza glabra: a study from the Hungry Steppes of Central Asia. International Journal of Agricultural Sustainability 3:102-113.
MAHONEY, J. M., AND S. B. ROOD. 1998. Streamflow requirements for cottonwood seedling recruitment--an integrative model. Wetlands 18:634-645.
NAIMAN, R. J., H. DECAMPS, AND M. E. MCCLAIN. 2005. Riparia: ecology, conservation, and management of streamside communities. Elsevier Academic Press, London, United Kingdom.
NAUMBURG, E., R. MATA-GONZALEZ, R. G. HUNTER, T. MCLENDON, AND D. W. MARTIN. 2005. Phreatophytic vegetation and groundwater fluctuations: a review of current research and application of ecosystem response modeling with an emphasis on Great Basin vegetation. Environmental Management 35:726-740.
NORD, E. A., AND J. P. LYNCH. 2009. Plant phenology: a critical controller of soil resource acquisition. Journal of Experimental Botany 60:1927-1937.
PASSIOURA, J. B. 2002. Soil conditions and plant growth. Plant Cell and Environment 25:311-318.
PERRY, L. G., D. C. ANDERSEN, L. V. REYNOLDS, S. M. NELSON, AND P. B. SHAFROTH. 2012. Vulnerability of riparian ecosystems to elevated CO2 and climate change in arid and semiarid western North America. Global Change Biology 18:821-842.
PLATT, D. R. 1988. Development and survival of plants in a prairie reconstruction at Kauffman Museum in south central Kansas. Professional Paper 09.02 in The prairie: roots of our culture, foundation of our economy, Proceedings of the Tenth North American Prairie Conference (A. Davis and G. Stanford, editors.). Native Prairie Association of Texas, Dallas.
QIAN, B., E. G. GREGORICH, S. GAMEDA, D. W. HOPKINS, AND X. L. WANG. 2011. Observed soil temperature trends associated with climate change in Canada. Journal of Geophysical Research-Atmospheres 116:D02106.
SCHADE, J. D., AND S. E. HOBBIE. 2005. Spatial and temporal variation in islands of fertility in the Sonoran Desert. Biogeochemistry 73:541-553.
SERRAJ, R., T. R. SINCLAIR, AND L. C. PURCELL. 1999. Symbiotic [N.sub.2] fixation response to drought. Journal of Experimental Botany 50:143-155.
SHAFROTH, P. B., G. T. AUBLE, AND M. L. SCOTT. 1995. Germination and establishment of the native plains cottonwood (Populus deltoides Marshall subsp. monilifera) and the exotic Russian-olive (Elaeagnus angustifolia L.). Conservation Biology 9:1169-1175.
SHAFROTH, P. B., A. C. WILCOX, D. A. LYTLE, J. T. HICKEY, D. C. ANDERSEN, V. B. BEAUCHAMP, A. HAUTZINGER, L. E. MCMULLEN, AND A. WARNER. 2010. Ecosystem effects of environmental flows: modelling and experimental floods in a dryland river. Freshwater Biology 55:68-85.
SHER, A. A., AND D. L. MARSHALL. 2003. Seedling competition between native Populus deltoides (Salicaceae) and exotic Tamarix ramosissima (Tamaricaceae) across water regimes and substrate types. American Journal of Botany 90:413-422.
SIVINSKI, R. C. 2007. Checklist of vascular plants in the Sandia and Manzano Mountains of central New Mexico. Occasional Papers of the Museum of Southwestern Biology 10:1-67.
VINSON, D. S., S. E. BLOCK, L. J. CROSSEY, AND C. N. DAHM. 2007. Biogeochemistry
at the zone of intermittent saturation: field-based study of the shallow alluvial aquifer, Rio Grande, New Mexico. Geosphere 3:366-380.
WOLDEN, L. G., J. C. STROMBERG, AND D. T. PATTEN. 1995. Flora and vegetation of the Hassayampa River Preserve, Maricopa County, Arizona. Journal of the Arizona-Nevada Academy of Science 28:76-111.
ZAHRAN, H. H. 1999. Rhizobium-legume symbiosis and nitrogen fixation under severe conditions and in an arid climate. Microbiology and Molecular Biology Reviews 63:968-989.
Submitted 17 July 2012. Acceptance recommended by Associate Editor Janis K Bush 29 March 2013.
DOUGLAS C. ANDERSEN * AND S. MARK NELSON
United States Geological Survey, Fort Collins Science Center, Centre Avenue, Building C, Fort Collins, CO 80526 (DCA) Bureau of Reclamation, 86-68220, P.O. Box 25007, Denver, CO 80225 (SMN)
TABLE 1--Results of pairwise analysis of growth curves for Glycyrrhiza lepidota (n = 26 time points) grown using four combinations of two soil temperatures (W = warm; C = cool) and two depths to the water table (L = large; S = small). Statistics are based on 4,000 permutations. Comparison Statistic P-value Adjusted P-value Treatment 1 Treatment 2 CL WL -0.611 0.41375 0.8275 CL CS -2.581 0.00100 0.0050 CL WS -3.785 0.00000 <0.0001 WL CS 1.745 0.02775 0.0833 WL WS -2.739 0.00175 0.0070 CS WS -0.616 0.49675 0.8275 TABLE 2--Results of two-way analysis of variance with a covariate on each of final biomass of shoots and roots and total biomass of Glycyrrhiza lepidota grown in mesocosms mimicking dryland riparian environments. Treatments involved two depths to ground water ([D.sub.W]) and two soil temperatures ([T.sub.S]). The covariate (position) was the sequentially numbered position of each stationary mesocosm along their linear arrangement in the greenhouse (Fig. 1). In all cases, n = 61. Response Source Sum of df Mean variable squares square Shoot [D.sub.W] 12.507 1 12.507 biomass [T.sub.S] 8.277 1 8.277 [D.sub.W] x [T.sub.S] 6.008 1 6.008 Position 19.086 1 19.086 Error 178.320 56 3.184 Root [D.sub.W] 138.675 1 138.675 biomass [T.sub.S] 16.814 1 16.814 [D.sub.W] X [T.sub.S] 17.671 1 17.671 Position 29.165 1 29.165 Error 1,038.710 56 18.548 Total [D.sub.W] 234.476 1 234.476 biomass [T.sub.S] 48.686 1 48.686 [D.sub.W] x [T.sub.S] 44.285 1 44.285 Position 95.437 1 95.437 Error 1,984.091 56 35.430 Response Source F-ratio P variable Shoot [D.sub.W] 3.928 0.052 biomass [T.sub.S] 2.599 0.113 [D.sub.W] x [T.sub.S] 1.887 0.175 Position 5.994 0.018 Error Root [D.sub.W] 7.476 0.008 biomass [T.sub.S] 0.906 0.345 [D.sub.W] X [T.sub.S] 0.953 0.333 Position 1.572 0.215 Error Total [D.sub.W] 6.618 0.013 biomass [T.sub.S] 1.374 0.246 [D.sub.W] x [T.sub.S] 1.250 0.268 Position 2.694 0.106 Error TABLE 3--Mean ([+ or -] 1 SE in parentheses) concentrations of nitrate (N[O.sub.3]-N) and ammonium (N[H.sub.4]-N), total nitrogen (N), total carbon (C), and C:N ratio of soil in pots at the end of an experiment on growth of Glycyrrhiza lepidota under variable conditions of soil temperature and depth to ground water. The level of N[O.sub.3]-N in soil prior to treatments was 10.0 mg-kg [+ or -] 0.0 SE (n = 3). Samples (n) collected post-treatment in pots with and without plants were each divided equally between two treatments of depth from the surface of the soil to the water table ([D.sub.W]; S = 63 cm and L = 100 cm). [D.sub.W] Pots with plants (n = 12) N[O.sub.3]- N[H.sub.4]- Total Total C:N N (mg/kg) N (mg/kg) N (%) C (%) ratio S 4.70 (a) 3.21 0.024 (b) 0.45 (b) 20.0 (0.27) (1.15) (0.003) (0.026) (2.09) L 1.92 (ab) 1.40 0.022 0.45 24.6 (0.29) (0.62) (0.004) (0.013) (4.56) [D.sub.W] Pots without plants (n = 6) N[O.sub.3]- N[H.sub.4]- Total Total C:N N (mg/kg) N (mg/kg) N (%) C (%) ratio S 4.87 (a) 0.87 0.049 (b) 0.57 (b) 12.6 (0.60) (0.32) (0.013) (0.04) (2.18) L 10.67 (ab) 0.80 0.038 0.38 9.0 (1.16) (0.15) (0.007) (0.14) (2.49) (a) Significantly different (P < 0.05) between [D.sub.W] = S and [D.sub.W] = L. (b) Significantly different (P < 0.05) between pots with and without plants.
|Printer friendly Cite/link Email Feedback|
|Author:||Andersen, Douglas C.; Nelson, S. Mark|
|Date:||Mar 1, 2014|
|Previous Article:||Documenting the early stages of invasion of Matthiola parviflora and predicting its spread in North America.|
|Next Article:||Lepidopteran herbivory in restored and successional sites in a tropical dry forest.|