Effect of Long-Term Semiarid Pasture Management on Soil Hydraulic and Thermal Properties

Semiarid pasture management strategies can affect soil hydraulic and thermal properties that determine water fluxes and storage, and heat flow in unsaturated soils. We evaluated long-term (>10 years) perennial and annual semiarid pasture system effects on saturated hydraulic conductivity (ks), soil water retention curves (SWRCs), soil water thresholds (i.e., volumetric water content (θv) at saturation, field capacity (FC), and permanent wilting point (PWP); plant available water (PAW)), thermal conductivity (λ), and diffusivity (Dt) within the 0–20 cm soil depth. Forage systems included: Old World bluestem (Bothriochloa bladhii) + legumes (predominantly alfalfa (Medicago sativa)) (OWB-legume), native grass-mix (native), alfalfa + tall wheatgrass (Thinopyrum ponticum) (alfalfa-TW), and annual grass-mix (annual) pastures on a clay loam soil; and native, teff (Eragrostis tef), OWB-grazed, and OWB-ungrazed pastures on a sandy clay loam soil. The perennial OWB-legume and native pastures had increased soil organic matter (SOM) and reduced bulk density (ρb), improving ks, soil water thresholds, λ, and Dt, compared to annual teff and alfalfa-TW (P < 0.05). Soil λ, but not Dt, increased with increasing θv. Grazed pastures decreased ks and water retention compared to other treatments (P < 0.05), yet did not affect λ and Dt (P > 0.05), likely due to higher ρb and contact between particles. Greater λ and Dt at saturation and PWP in perennial versus annual pastures may be attributed to differing SOM and ρb, and some a priori differences in soil texture. Overall, our results suggest that perennial pasture systems are more beneficial than annual systems for soil water storage and heat movement in semiarid regions.


Introduction
Soil hydraulic properties, which are needed for analyzing water fluxes and storage in unsaturated soils, can play an essential role in agricultural water management, especially in water-scarce regions [1]. Broadly, soil hydraulic properties, such as saturated hydraulic conductivity (k s ) and soil water retention characteristics or curve (SWRC), are affected by both inherent and dynamic soil properties that can be managed or controlled, at least to a certain extent [2]. These soil properties include soil texture, pore size distribution, mineralogical composition, soil structure and aggregation, infiltration, bulk density (ρ b ), and soil organic matter (SOM) content [3,4]. These properties can be influenced by various factors, including land uses [5,6], root biomass [7], soil fauna [8], shrinking and swelling properties of clay soils [9], freezing and thawing [10,11], and agricultural activities such as tillage [12], wheel-traffic compaction [13], and animal trampling [14].
Similarly, soil thermal properties influence the partition of energy at the soil surface and control the soil thermal regime, thereby affecting coupled liquid water, water vapor, effect of these pasture systems on soil thermal properties as a function of soil water contents given by soil water retention curves (SWRCs).

Study Site
The study was conducted at the Texas Tech University's Forage-Livestock Research Laboratory Farm, New Deal, located 24 km northeast of Lubbock, Texas, USA (33°43′47.2″ N, 101°43′47.1″ W at 993 m above mean sea level). The study site represented unsaturated soils that have undergone semiarid pasture management since 1997 [22]. This included variable precipitation conditions typical of the Texas Southern High Plains and various forage crop species, cattle grazing systems, water management practices (i.e., irrigation and dryland), and tillage methods. The land was nearly level, with 0 to 1% slopes. The taxonomic class of soil was Pullman clay loam (fine, mixed, superactive, thermic Torrertic Paleustolls), extensive in MLRA (Major Land Resource Area) 77C of the Southern High Plains [32]. The soil has 35 to 50% silicate clay and 0 to 3% carbonate clay content [32]. The diagnostic horizons include Ap, Bt1, Btk1, Btk2, and Btk3, with depth to secondary carbonates at 50 to 76 cm and calcic horizon at 76 to 150 cm [32]. The site has a continental, semiarid climate where the long-term mean annual precipitation for Lubbock County is 469 mm, and annual evapotranspiration is 1500 mm [33]. Monthly precipitation and air temperatures are presented in Figure 1. During the study period (2016-2017), 2016 was drier than 2017 when compared to the long-term average precipitation, especially during summer (June-September) and winter (November-December). The average daily maximum temperature exceeded 30 °C from June to August, while the minimum temperature remained below 0 °C from December to January.

Experiment Design and Pasture Management
The experimental site for this study had two major pasture areas, of which the east pasture area (12.3 ha) (Figure 2, top) was smaller than the west pasture (32.7 ha) ( Figure  2, bottom). The establishment and management of these pastures were described by [22] and [34] (Supplementary Tables S1 and S2). The east pasture area was established in 1983 and was under perennial forages for many years (>30 years). The west pasture area, established in 2002, was previously under continuous cotton (Gossypium hirsutum L.)

Experiment Design and Pasture Management
The experimental site for this study had two major pasture areas, of which the east pasture area (12.3 ha) (Figure 2, top) was smaller than the west pasture (32.7 ha) (Figure 2, bottom). The establishment and management of these pastures were described by [22] and [34] (Supplementary Tables S1 and S2). The east pasture area was established in 1983 and was under perennial forages for many years (>30 years). The west pasture area, established in 2002, was previously under continuous cotton (Gossypium hirsutum L.) system. Treatments were only compared within, rather than between, these selected pasture areas to avoid confounding effects of time since the establishment and management history. The differences in ρ b , soil texture, and SOM between these pastures were also emphasized in previous studies [35,36]. Over the years, the establishment and management practices, such as water and irrigation management, crop rotation, use of various forage species, grazing, and tillage methods, were different for each pasture type or system in both east and west pasture areas. Different pasture systems were categorized into four broad practices (or treatments) in each pasture area to evaluate hydraulic and thermal properties in the upper 0-20 cm soil depths under long-term pasture management practices. In addition, a comparison was made between two specific pasture treatments: a grass-only (monoculture or grass-mix) and a grass-legume (two or more species) mixture. system. Treatments were only compared within, rather than between, these selected pasture areas to avoid confounding effects of time since the establishment and management history. The differences in ρb, soil texture, and SOM between these pastures were also emphasized in previous studies [35,36]. Over the years, the establishment and management practices, such as water and irrigation management, crop rotation, use of various forage species, grazing, and tillage methods, were different for each pasture type or system in both east and west pasture areas. Different pasture systems were categorized into four broad practices (or treatments) in each pasture area to evaluate hydraulic and thermal properties in the upper 0-20 cm soil depths under long-term pasture management practices. In addition, a comparison was made between two specific pasture treatments: a grass-only (monoculture or grass-mix) and a grass-legume (two or more species) mixture. During the study period (2016-2017), the west pasture area had pasture systems that contained either teff (Eragrostis tef (Zucc.) Trotter), native grass-mix (buffalograss (Buchloe dactyloides (Nutt.) Engelm), blue grama (Bouteloua gracilis (Willd. Ex Kunth) Lag. Ex Griffths), sideoats grama (Bouteloua curtipendula (Michx.) Torr), and green sprangletop (Leptochloa dubia (Kunth.) Nees)), or Old-World bluestem cv. WW-B.Dahl (OWB, Bothriochloa bladhii (Retz) T. Blake). The east pasture area had pasture systems containing either mixtures of OWB + alfalfa (Medicago sativa L.) + yellow sweetclover (Melilotus officinalis L.), alfalfa + tall wheatgrass (TW, Thinopyrum ponticum (Host) Beauv.), native, or teff (Eragrostis tef) (only in 2016). Since one of the replicate pastures in the east area had teff only in 2016 and was under different annual crops in previous years, such as sorghum-sudangrass (Sorghum bicolor (L.) Moench), wheat (Triticum aestivum L.), cereal rye (Secale cereale L.), or left fallow (annual weeds) during a severe drought in 2011 and 2012, the pasture was treated as annual. The native pastures in the east area did not have buffalograss. The grazing versus ungrazed (haying only) effect on soil hydraulic and thermal properties were evaluated for OWB pastures in the west area.
The experiment compared treatments using errors estimated with three replicate pastures nested within each pasture area ( Figure 2). The east area replicates contained forage types that composed the grass-only and grass-legume systems to compare. A more favorable design would have all treatments within each spatial replicate; nevertheless, permanent physical structures of subsurface irrigation systems, fencing, and established stands did not allow the redesign of treatments and replicates. This study, aimed at evaluating the effect of perennial and annual pasture systems on soil hydraulic and thermal properties, was laid out in accordance with the layout of pasture treatments carried out over the years in the east and west pasture areas. The treatment arrangement in this study may have incurred unforeseen errors owing to a potential confounding of natural differences in soil properties between the east and west areas, for example, depth to the CaCO 3 layer. However, the north-south placement of replicates and the separation of the east area from the west area may have minimized the potentially confounding effects on treatment differences.

Sample Collection and Preparation
Soil sampling was carried out from September to November in 2016 and from March to May in 2017. Soil samples were collected from three spatial points stratified in each pasture ( Figure 2; see triangles in the west native treatment). Samples were taken from each end (10 m from the fence) and the middle of the pastures while maintaining a distance of at least 15 m between the sampling points. Both bulk and core soil samples were collected from each sampling point. Bulk samples were collected from 0-10 cm and 10-20 cm depths using a shovel in each treatment plot. The vegetative cover was not considered part of the surface soil sample. Collected bulk samples were stored in polyethylene zipper bags, labeled, and placed in a container to avoid evaporation. A total of 198 bulk samples (11 pastures × 3 replicates × 3 locations × 2 soil depths) were collected from the experimental site each year.
Similarly, undisturbed core samples were collected using 5 × 5 cm cylindrical cores at each sampling point. Core sampling was carried out with the help of a core sampler (with slide hammer attachment) at different soil depths down to 20 cm (i.e., 0-5, 5-10, 10-15, and 15-20 cm). Each aluminum liner with a core sample was removed from the core sampler, labeled, and sealed immediately with polyethylene end caps at both ends of the liner to prevent both soil loss and evaporation. A total of 396 soil cores (11 pastures × 3 replicates × 3 locations × 4 soil depths) were collected from the experimental site each year. All soil samples were transferred from the experimental site to the Soil Physics Laboratory at Texas Tech University. Bulk samples were air-dried, crushed, and sieved to pass a 2 mm screen before further analysis of soil properties. Bulk and core samples were stored in the refrigerator at 4 • C before analysis.

Basic Soil Properties
The gravimetric water content (θ g ) of soil samples was determined from a 55 g subsample taken from fresh bulk samples. A pre-weighed aluminum foil was used to place the soil in a drying oven (Thermo Fisher Scientific, Waltham, MA) at 105 • C for at least 24 h until a constant dry weight was recorded. The θ g (g g −1 ) was determined by dividing the weight of water by the weight of dry soil. About 50 g of air-dried sample (<2 mm) was used to conduct particle size analysis using the sedimentation (hydrometer) method [37] with the ASTM 152H hydrometer. Hydrometer readings and suspension temperature data were taken at 40 s and after 3 h, which were then used to estimate sand, clay, and silt percentages. The 5 × 5 cm core samples were used to determine soil ρ b (g cm −3 ) by the core method [38], in which ρ b was calculated as the ratio of the mass of dry soil solids (after oven drying at 105 • C for 24 h) to the bulk volume (volume of solids plus soil pores). The SOM content of soil samples was determined using the loss-on-ignition procedure that estimates SOM based on the gravimetric weight change associated with the high-temperature oxidation of organic matter [39]. About 10 g bulk soil sample was dried in an oven at 105 • C for 24 h and then ignited in a muffle furnace (Thomas Scientific, Swedesboro, NJ, USA) at 400 • C for 16 h. The difference in weight before (i.e., the weight of oven-dried soil at 105 • C) and after ignition (i.e., the weight of soil at 400 • C) was taken as the amount of SOM (mg g −1 ) that was present in the sample.

Soil Hydraulic Conductivity and Soil Water Retention Characteristics
Measurements of k s (cm d −1 ) were made on soil core samples (8 cm in diameter and 5 cm in height) by using both constant head and falling head methods [40] with a KSAT benchtop instrument (Meter Group, Inc., Pullman, WA, USA). Before k s measurement, air-dried bulk soil samples from 0-10 and 10-20 cm depths with known ρ b values were repacked into 8 × 5 cm sampling cylinders. The measurement of k s was repeated three times for each core sample. In addition, a limited number of core samples using instrumentspecific 8 × 5 cm cylinders were also collected from 0-10 and 10-20 cm soil depths to verify k s measurements of saturated repacked soil cores with saturated intact cores using the KSAT instrument. Both constant head and falling head methods, which determine soil k s using Darcy's equation, are represented by the following governing equations (Equations (1) and (2), respectively).
where k s is the saturated hydraulic conductivity (cm s −1 ), V (cm 3 ) is the volume of water flowing through the cross-sectional area A (cm 2 ) of the soil sample per unit time t (s), Q is the steady-state flow rate from Mariotte flask (cm 3 s −1 ), H is the hydraulic head difference between the water inlet and outlet level (cm), L is the length of the soil sample (cm), a is the cross-sectional area of the burette (cm 2 ), H 0 and H t are the initial and final hydraulic head differences in the burette (cm), and t t is the time taken for the pressure head to drop from the initial to final pressure head in the burette (s). Soil water retention characteristics were determined using the pressure chamber method [41] for all treatments in each pasture area using 5 × 5 cm core samples taken at soil depths of 0-5, 5-10, 10-15, and 15-20 for each year. A pressure plate apparatus (Soil Moisture Equipment Corp., Santa Barbara, CA, USA) was used to determine SWRC at nine pressure heads (h): 0 (at saturation), −100, −330, −500, −2000, −3000, −5000, −10,000, and −15,000 cm. Core samples were saturated from the bottom, weighed, and then placed on a saturated ceramic pressure plate in the pressure plate extractor. A predetermined h was applied for 36 to 48 h. Once the cores were equilibrated to applied h, cores were removed from the extractor and weighed. The same procedure was repeated with successively lower (more negative) h (i.e., from −100 to −15,000 cm). The ceramic plates corresponding to higher h (saturation to −3000 cm) and lower h (−5000 to −15,000 cm) were used to determine SWRCs (soil water retention curves) in the wet and dry soil water ranges, respectively. At the end of the experiment, gravimetric (θ g ) and volumetric (θ v ) water contents (cm 3 cm −3 ) of the core samples were estimated from known ρ b (g cm −3 ) of the core samples, dry weight of the core samples, and weight of the core samples at the end of each pressure head (h) experiment. The θ v determined at each applied h was averaged over three replicate soil cores. The θ v values were plotted against the corresponding h values to obtain functional relationships between θ v and h of the soil samples, i.e., SWRCs for all treatments in the east and west pasture areas.
To quantify water flow parameters in unsaturated soils under long-term pasture management practices in both pasture areas, the van Genuchten's SWRC model [42], which has shown its feasibility for a wide variety of soils (e.g., [43,44]), was fitted to measured SWRCs at different soil depths (i.e., 0-5, 5-10, 10-15, and 15-20 cm) for all treatments, using a nonlinear least-squares optimization program RETC [43]. where θ r and θ s are the residual and saturated θ v , respectively (cm 3 cm −3 ); h is the pressure head (cm); and α v (is related to the inverse of air entry pressure head; cm −1 ) and n (is related to pore size distribution; dimensionless) are fitting shape parameters of the SWRC. Soil water thresholds, necessary for water management in agricultural systems, were obtained from measured SWRCs at different soil depths for all treatments in each pasture area. Soil water for h > −15,000 cm was considered unavailable to plants. The θ s (i.e., θ v at saturation) provided a measure of the total porosity of soil samples. Plant available water (PAW) was determined as the difference between field capacity water content (FC; θ v at h = −330 cm), which is the upper limit of PAW, and permanent wilting point (PWP; θ v at h = −15,000 cm) when the plant can no longer extract water from the soil.

Soil Thermal Properties
For all treatments in each pasture area, measurements of λ , and D t (mm 2 s −1 ) were made on 5 × 5 cm core samples taken at soil depths of 0-5, 5-10, 10-15, and 15-20 cm using the KD2 Pro Thermal Properties Analyzer (Meter Group, Inc., Pullman, WA, USA). The λ, C v , and D t values were measured simultaneously during the SWRC measurements at applied pressure heads (h) of 0 (at saturation), −100, −330, −500, −2000, −3000, −5000, −10,000, and −15,000 cm. The KD2 pro instrument, equipped with a dual-needle sensor (30 mm long, 1.3 mm diameter, and 6 mm spacing), works on principles of the transient line heat source analysis [45,46]. The instrument was calibrated before measuring λ, C v , and D t of core samples using a Delrin verification block supplied by the manufacturer.
During SWRC measurements, after the soil water equilibrium had been reached at each of the applied pressure heads, the core samples were removed from the pressure extractor, and the dual-needle sensor was inserted at the full depth of the needles into each core sample to determine λ, C v , and D t . During the measurement of λ, C v , and D t , the KD2 pro instrument applied heat for a set of heating times (t h ) to one of the needles that contained a heating element. The heat was then transferred into the soil between two needles, and the temperature was measured in the measuring needle during the heating and the cooling period following heating. The resulting data were fitted to the following equations (Equations (4) and (5)) using a nonlinear least-squares procedure [47].
where ∆T is the temperature rise at the measuring needle; T 0 is the temperature at the start of the measurement; T is the measured temperature; b 0 , b 1 , and b 2 are the fitting parameters; q is the heat input at the heated needle (W m −1 ); E i is the exponential integral that is approximated using polynomials [48]; t is time (s); and t h is the heating time (s). Equation (4) was applied for the first t h when the heat was on, while Equation (5) was applied when the heat was off. The λ and D t were then computed from Equations (6) and (7), respectively, while C v was given as the ratio of λ and D t .
where r is the distance from the heated needle to the measuring needle.

Statistical Analysis
Measured data were analyzed separately for the east and west pasture areas. The generalized linear mixed model (PROC GLIMMIX) procedure in SAS version 9.4 [49] was used [50] to test the statistical significance of measured soil hydraulic and thermal properties among pasture treatments, soil depths, treatment × depth interactions, and depth × year interactions within each pasture area. Treatment differences were considered significant when ≥least significant difference (LSD) at α = 0.05. Pasture treatments were set as a fixed effect, whereas replication was set as a random effect. The repeated measurement analysis was performed for changes over the years [51], in which the denominator df (degrees of freedom) was adjusted to obtain an appropriate standard error using the Kenward-Roger method [52]. The relationship between variables was analyzed using a regression procedure (PROC REG). The outputs from the RETC model were exported to SigmaPlot 12.5 [53] to draw SWRCs for each treatment and depth.

Basic Soil Properties
Sand, silt, and clay contents in unsaturated soils were not affected by management practices in both east and west pasture areas (P > 0.05, Table 1). The east area was more clayey (i.e., clay loam according to the USDA classification) than the west (i.e., sandy clay loam). The variability in soil texture observed in pasture areas was consistent with previous studies [36,54]. SOM contents differed among treatments at 0-20 cm depth (Table 1). There was a treatment × year interaction for the east area (P < 0.01), while treatments did not affect SOM in the west in 2017 (P > 0.05). In the east pasture area, OWB-legume had the greatest SOM compared to alfalfa-TW, native, and annual (P < 0.01), whereas the annual had the lowest SOM (P < 0.01). OWB-grazed and OWB-ungrazed had consistently greater SOM than teff in 2016 (P < 0.05), while the native was intermediate between OWB and teff. Grazing did not affect SOM in the west OWB pasture (P > 0.05).
The greater SOM content at OWB pastures could be attributed to its vegetative characteristics. It has been evident that OWB-legume produced more biomass than native, alfalfa-TW, teff, and other annual grasses because of its ability to tolerate drought and longer vegetative period [34,55,56]. This suggested that more organic matter could return to the soil after the decomposition of aboveground residue and roots. Various OWB pasture systems have been reported to enhance microbial biomass [57] and soil organic C [58].
OWB pasture in the east had greater SOM than in the west (Table 1), which was most likely the result of the inclusion of legume and perennial species over more years. In a previous study conducted by Bhandari et al. (2018) [35], OWB-alfalfa pastures had greater soil organic C than native pastures in the east area, owing to a larger microbial population size. Generally, grazing animals help mix plant residues and manure into the soil, thereby increasing microbial substrate quality. However, responses of microbial biomass C and microbial community structure to grazing have yielded contrasting results in different studies (e.g., [57,[59][60][61] [60] reported lower microbial biomass in grazed pasture soils compared to ungrazed soils. Overall, our results indicated that OWB, a C4 perennial grass, was very effective in building SOM, while annuals and natives were less effective. In both pasture areas, ρ b values were lower in the upper 0-5 cm soil depth (1.41 and 1.44 g cm −3 for the east and west areas, respectively) than in the 5-20 cm depth ( Table 1). The treatment × year interaction was not significant for both pasture areas (P > 0.05). In the east area, ρ b in the 0-5 cm depth differed among pasture treatments (P < 0.05), whereas treatment had no effect in the deeper soil depths (P > 0.05). OWB-legume pasture had lower ρ b than alfalfa-TW and annual in the upper 0-5 cm, while native had intermediate ρ b values between OWB-legume and alfalfa-TW and annual pastures. On the contrary, ρ b in the 0-5 cm was not affected by treatments in the west area (P > 0.05). Nonetheless, the treatment effect on ρ b was observed for 5-10 and 10-15 cm depths (P < 0.05). Ungrazed OWB showed lower ρ b than OWB-grazed and teff (P < 0.05). As in the east pasture, natives had an intermediate ρ b value between OWB and teff in the west area. Grazing increased ρ b in OWB pasture by 10% over OWB-ungrazed within the 5 to 15 cm soil depth in the west area, likely owing to animal trampling as SOM content was the same in both OWBgrazed and OWB-ungrazed pastures (Table 1). Animal trampling might decrease soil macroporosity (i.e., large pore spaces) and the total pore volume of soil, thereby increasing ρ b . Higher ρ b values under annual and teff pasture systems could be attributed to wheel traffic compaction and destruction of macropores due to tillage during planting. During soil sampling, the average gravimetric water content of the top 20 cm of soil was 10% and 11% in the east and west areas, respectively (Table 1). Native in 2016 and OWB-legume in 2017 in the east area had more soil water content than alfalfa-TW and annual. During soil sampling in the west pasture area in both years, OWB-grazed and OWB-ungrazed had more soil water content than native and teff (Table 1). Table 1. Basic soil properties in the east and west pasture area of the experimental site. Bulk density (ρ b ) and particle size distribution (soil texture) data were averaged across years. Soil texture, soil organic matter (SOM), and gravimetric water content (θ g ) were determined in the 0-20 cm soil depth.

Soil Hydraulic Properties and Water Retention Characteristics
Saturated hydraulic conductivity (k s ) values were 61 and 57 cm d −1 at depths of 0-10 and 10-20 cm, respectively, in the west pasture area during 2016-2017. The respective k s values for these depths in the east area were greater, i.e., 113 and 103 cm d −1 , even though the east pasture soils had higher clay contents. Notably, the east pasture area tended to have lower ρ b values than the west (Table 1). Hence, the most plausible explanation for differences in k s between pasture areas was that changes in ρ b and the resultant changes in total soil porosity under different long-term pasture management practices comprised the primary factors influencing soil k s . Total porosity (i.e., θ v at saturation determined from SWRCs) and characteristics of pores in soils have been known to impact k s . In addition, changes in vegetation species could be a driving factor influencing hydraulic properties by altering total porosity, non-capillary porosity, and macro water-stable aggregates [62]. Different vegetation types have also been reported to affect k s through root distribution and morphological characteristics, such as root biomass and distribution (e.g., [62][63][64][65]).
As shown in Figure 3, k s in the upper 0-10 cm soil depth was not affected by pasture management in the east area (P > 0.05), while it was significant for the 10-20 cm depth in the east and both 0-10 and 10-20 cm depths in the west area (P < 0.05). In the east area, native pasture had higher k s than annual and alfalfa-TW in the 10-20 cm depth (P < 0.05). For the 0-10 cm depth in the west area, OWB-ungrazed had higher k s than teff, and teff had the lowest k s (P < 0.05). Native and OWB-grazed pastures had intermediate k s values between OWB-ungrazed and teff. In the west area, k s in the 10-20 cm was in the order of native > OWB-ungrazed > OWB-grazed > teff. Saturated hydraulic conductivity (k s ) was associated with ρ b (r = 0.61, P < 0.0001) and SOM (r = 0.69, P < 0.0001) (data not shown) as it was greater for perennial pastures with low ρ b and high SOM than annual. Our result was similar to previous studies, in which lower ρ b was aligned with higher k s (e.g., [62,66,67], and k s tended to decrease with an increase in ρ b with soil depth [24,62]. the air entry h required to initiate the release of water and air begins to enter the soil pores during drainage when h exceeds air entry h, and especially parameter n, which describes the rate of water desorption, varied among pasture treatments within the 0-20 cm soil depths ( Table 2). The variations in αv and n were likely due to the effect of pasture management practices on soil micropores. Soil water thresholds, determined using SWRCs data (i.e., different soil water states from saturation to PWP) at different soil depths for each treatment, are shown in Table 3. Broadly, soil water thresholds were not affected by treatment × soil depth, year × treatment, and soil depth × year interactions at α = 0.05. An important aspect to be noted in Table 3 is that the perennial pastures, such as OWB and native, had improved soil water retention characteristics compared to annual teff (P < 0.05), with greater FC (0.38 to 0.44 cm 3 cm −3 ) and PAW (0.17 to 0.23 cm 3 cm −3 ). Generally, the east area had more water holding capacity (and PAW) than the west, which could be explained by the higher clay contents and SOM of the east pasture soils (Table 1).
It is worth noting that the type of OWB species used in this experiment (i.e., WW-B. Dahl) has been promoted for its longer vegetative period than other OWB species, which favors greater production of above-and belowground biomass [55]. Although the relation between SWRC or ks and root characteristics (e.g., distribution and morphological charac- The van Genuchten model was fitted to the measured SWRCs for different soil depths down to 20 cm to provide an additional tool for reliable estimates of SWRC parameters (θ s , θ r , α v , and n) for all treatments (Figure 4). Generally, the overall fit resulted in close agreements between measured and estimated θ v , as indicated by lower values of overall RMSE (root mean squared error) between 0.010 and 0.038 cm 3 cm −3 and 0.013 and 0.025 cm 3 cm −3 (data not shown) for the east and west pasture areas, respectively. The measured and fitted SWRCs for different pasture soils suggested that soils with greater SOM and k s retained more water at or near saturation (i.e., air-entry region) and released less water when the h increased (i.e., more negative h), except for native and OWB-ungrazed pastures in the west area. Nonetheless, the capillary region (i.e., after air began to enter) of the SWRC was steeper with an increase in h for soils with greater SOM (Figures 3 and 4; Table 1).   Similar to measured SWRCs at different soil depths, the patterns predicted by the modeled van Genuchten SWRCs differed among pasture treatments within and between the east and west areas ( Figure 4). Accordingly, the van Genuchten parameters, especially θ s , α v , and n, showed differences as a result of pasture management practices, thereby resulting in variations in soil water thresholds (i.e., θ v at saturation, FC, PWP, and PAW) in the wet and dry soil water ranges of SWRCs. Fitted van Genuchten parameter values within the 0-20 cm soil depth for all treatments in both pasture areas are listed in Table 2. The model fitting process yielded similar θ r values among different treatments in both pasture areas. The residual water content (θ r ) represents θ v at the lowest h, at which water is retained in soil micropores [42,43,70]. The θ r is often related to the amount and state of water adsorption in the mineral fractions of soil [71], which might not be affected by different pasture management practices. On the contrary, parameter α v , which determines the air entry h required to initiate the release of water and air begins to enter the soil pores during drainage when h exceeds air entry h, and especially parameter n, which describes the rate of water desorption, varied among pasture treatments within the 0-20 cm soil depths ( Table 2). The variations in α v and n were likely due to the effect of pasture management practices on soil micropores. ; θ s , volumetric water content at saturation (cm 3 cm −3 ); α v (related to the inverse of air entry pressure head; cm −1 ) and n (related to pore size distribution; dimensionless) are fitting shape parameters. Abbreviations: OWB, Old-World bluestem; TW, tall wheatgrass.
Soil water thresholds, determined using SWRCs data (i.e., different soil water states from saturation to PWP) at different soil depths for each treatment, are shown in Table 3. Broadly, soil water thresholds were not affected by treatment × soil depth, year × treatment, and soil depth × year interactions at α = 0.05. An important aspect to be noted in Table 3 is that the perennial pastures, such as OWB and native, had improved soil water retention characteristics compared to annual teff (P < 0.05), with greater FC (0.38 to 0.44 cm 3 cm −3 ) and PAW (0.17 to 0.23 cm 3 cm −3 ). Generally, the east area had more water holding capacity (and PAW) than the west, which could be explained by the higher clay contents and SOM of the east pasture soils (Table 1).  It is worth noting that the type of OWB species used in this experiment (i.e., WW-B. Dahl) has been promoted for its longer vegetative period than other OWB species, which favors greater production of above-and belowground biomass [55]. Although the relation between SWRC or k s and root characteristics (e.g., distribution and morphological characteristics) was not clarified in our study, root parameters might have had effects on SWRC or k s under different pasture systems [62,72]. In the east area, θ v at saturation (i.e., total soil porosity) in the 0-5 cm depth was greater for OWB-legume than alfalfa-TW and annual (P < 0.02), suggesting that OWB-legume treatment provided the relatively good topsoil structure.
As shown in Figure 4, when compared with native, annual, and alfalfa-TW, OWBlegume pasture retained nearly 0.05 and 0.1 cm 3 cm −3 more water content in the 0-5 cm between h = 0 cm (at saturation) and h = −100 cm. Alfalfa-TW had the lowest total porosity in the 5-10 cm, where there was a parity among OWB-legume, native, and annual treatments (P < 0.05) ( Table 3). In general, OWB-legume treatment showed greater soil water retention up to h of −2000 cm, while all pasture treatments, except native, had similar soil water retention characteristics after h of −2000 cm. Alfalfa-TW showed poor soil water retention up to h of −2000 cm (Figure 4). Total porosity at soil depths of 10-15 and 15-20 cm was not different among pasture treatments in the east area.
In the west pasture area, teff provided lower θ v at saturation than OWB and native perennial pastures (P < 0.05) in the upper 0-5 cm depth (Table 3; Figure 4). As shown in Figure 4, native and both grazed and ungrazed OWB tended to have greater soil water retention in the 0-5 cm at h less than −100 cm. At the 5-10 cm depth, soil water retention was in the order of native > OWB-ungrazed > OWB-grazed > teff up to h of −330 cm (i.e., FC), but after that, native and teff released more water at an increasing rate with increasing (i.e., more negative) h when compared with OWB pastures. OWB-grazed delayed release of water after h of −1000 cm. OWB-ungrazed retained water more tightly, close to the grazed one in the 10-15 cm under h of −2000 cm or lower. At the 15-20 cm depth, native and OWB-ungrazed had improved soil water retention characteristics at h less than −100 cm than teff and OWB-grazed. Nevertheless, native and OWB-ungrazed released water under increasing h, while the soil impacted by grazing (OWB-grazed) held more water under the increased (i.e., more negative) h (Figure 4). Grazing reduced total porosity (P < 0.05) in the 15-20 cm, which was suggested by higher ρ b at this 15-20 cm soil depth.
Similar effects of pasture management to those discussed for θ v at saturation were observed for FC soil water threshold (i.e., θ v at field capacity) ( Table 3). In the east pasture area, OWB-legume and native had greater FC than alfalfa-TW (P < 0.05) in the 0-5 cm. Annual pasture had an intermediate FC value between native and alfalfa-TW. Native in the 5-10 cm and OWB-legume in the 10-20 cm had greater FC than alfalfa-TW (P < 0.05). Both native and annual treatments provided mostly similar intermediate FC values between OWB-legume and alfalfa-TW. In the west area, the effects of pasture management on FC soil water thresholds were not significant for 0-5 and 5-10 cm depths (P > 0.05). However, the effects of pasture management on FC for 10-15 and 15-20 cm depths were significant (P < 0.05). OWB-grazed treatment provided lower FC than OWB-ungrazed and native at depths of 10-20 cm and 15-20 cm, respectively (P < 0.05). As mentioned earlier, although soils in the west area had higher sand contents, higher ρ b under OWB-grazed likely diminished the volume of large pores, affecting soil water retention at field capacity in denser soils.
As compared to FC soil water thresholds, an opposite trend was observed for PWP soil water threshold values under different treatments (Table 3), where pasture systems did not affect PWP in the east area (P > 0.05). In the west area, OWB-grazed showed greater PWP than OWB-ungrazed and native in the upper 0-5 cm (P < 0.05). OWB-grazed showed higher PWP compared to native pasture (P < 0.05) in the 0-5 cm depth (P < 0.05). Notably, grazing in the west area did not affect θ v at PWP at soil depths of 5-10 cm and 15-20 cm (P > 0.05).
Following the differences observed in FC and PWP soil water thresholds at different soil depths, PAW ranged from 0.112 to 0.233 cm 3 cm −3 and 0.095 to 0.190 cm 3 cm −3 for the east and west pasture areas, respectively, within the 0-20 cm depth ( Table 3). The east area pasture systems on clayey soils, associated with greater SOM, showed more PAW than the west. This was in alignment with prior studies (e.g., [73,74]), in which the increases in FC and PWP, as well as the greater available water content, were found for soils with higher clay and silt contents. OWB-legume had the highest PAW in the east area in the upper 0-5 cm compared to native, annual, and alfalfa-TW (P < 0.05). However, for the 5-10 cm depth, native pasture showed greater PAW than OWB-legume (P < 0.05). Alfalfa-TW treatment had PAW (0.120 cm 3 cm −3 ), almost half of that in the OWB-legume treatment (0.229 cm 3 cm −3 ). In the west pasture area, native and OWB-ungrazed had greater PAW than OWB-grazed in all soil depths (P < 0.05).

Soil Thermal Properties
Soil thermal conductivity (λ) and diffusivity (D t ) as a function of volumetric water contents (θ v ), which correspond to applied h of 0 (at saturation),−100, −330, −500, −2000, −3000, −5000, −10,000, and −15,000 cm, are shown in Figure 5 for all treatments at upper soil depths from 0 to 20 cm. Soil λ increased with increasing θ v in both pasture areas. The increase in λ values with increasing θ v agreed with previous studies (e.g., [75]). Soil λ was maximum at saturation and minimum at PWP. In dry soils at PWP (<0.05 cm 3 cm −3 ; Figure 5), as the contact points between solid particles are very small compared to the contact points between air and solid particles, heat transfer is governed by conduction within the air and by heat transfer across the gas-solid interface. Hence, λ of dry soils is primarily controlled by the gaseous phase and is usually low owing to the lower thermal conductivity of air than that of the other soil components (i.e., λ air < λ water < λ mineral ). As θ v increases towards saturation, the replacement of air with water continues to enhance heat conduction through the mixture because more water gathers around the contact points and forms water bridges (i.e., thermal bridges) between solid particles [76,77]. Notably, although λ increased with increasing θ v , there were no discernible changes in D t with increasing θ v for all treatments in both pasture areas. The reason could be explained by very little relative changes in the λ and C v values of pasture soils observed in our study. The increase in λ with increasing θ v was lower than the increase in C v . On average, C v increased by 13 and 7% when θ v increased from PWP (2.33 and 2.35 MJ m −3 K −1 ) to saturation (2.63 and 2.51 MJ m −3 K −1 ) (data not shown) for the east and west pasture areas, respectively. This also suggested that while higher λ values ensured faster soil temperature recovery, higher C v values of pasture soils likely provided larger heat storage in the upper 0-20 cm depth.
As shown in Figure 5, D t values (as a function of θ v ) varied between 0.39 and 0.80 mm 2 s −1 and between 0.48 and 0.76 mm 2 s −1 within the 0-20 cm soil depth in the east and west pasture areas, respectively. Soil D t data observed in our study were within the desirable ranges reported in previous studies [16,78,79], suggesting that both pasture soils would rapidly adjust to any temperature changes or would prevent the occurrence of temperature extremes when exposed to changes in the thermal environment. As reported by several studies (e.g., [20,80]), λ values in our study were also within and close to the desirable λ ranges that facilitated greater heat movement through the soil. For instance, Ghuman and Lal (1985) [20] found that λ ranged from 0.37 to 1.42 W m −1 K −1 at θ v of 0.02 to 0.16 cm 3 cm −3 for sandy loam, 0.35 to 3.34 W m −1 K −1 at θ v of 0.02 to 0.46 cm 3 cm −3 for sandy clay loam, and 0.39 to 1.15 W m −1 K −1 at θ v of 0.10 to 0.52 cm 3 cm −3 for clay soil, respectively.
The depthwise comparison of λ values for each pasture area showed a sudden increase in λ when θ v increased from 0.20 to 0.25 cm 3 cm −3 , after which it increased slowly with further increases in θ v up to saturation ( Figure 5). The maximum λ of 2.03 W m −1 K −1 was observed at saturation in the 5-10 cm depth under the native pasture system. The regression between λ and θ v ( Figure 6) showed similar slopes of the linear function with a higher y-intercept for the west pasture soils than the east. In the east pasture area, native and alfalfa-TW had greater λ than OWB-legume at both saturation and PWP in the upper 0-5 cm depth (P < 0.05), while λ value of annual pasture soils was intermediate. Except for alfalfa-TW, the differences in D t at saturation among native, alfalfa-TW, and OWB-legume followed a similar pattern of λ values for the same soil depth. For the 5-20 cm depth, D t at saturation did not vary between treatments. At PWP, native had greater D t than OWB-legume and alfalfa-TW in the 0-5 cm, whereas, in the 5-10 cm depth, OWB-legume and annual had greater D t than alfalfa-TW (P < 0.05).
Despite having lower θ v (a function of pressure heads) in the west pasture soils (Figure 4), the higher λ and D t in the west compared to the east could be explained by various factors, such as soil texture or particle size distribution, SOM, and ρ b [81], which were affected by pasture management practices. Soils with higher clay contents in the east pasture (i.e., clay loam soils) likely possessed a higher degree of aggregation than soils in the west pasture (i.e., sandy clay loam soils) with greater sand contents. This suggested that poorly aggregated soils in the west area had smaller pore spaces with better contact between soil solid particles than well-aggregated soils in the east. The results agreed with previous studies, in which lower λ values were observed for clayey soil compared to sandy soil (e.g., [75,82,83]). The relatively lower ρ b in the east pasture area could contribute to greater pore space, resulting in a decrease in λ and D t [20,83]. On the contrary, the contact between individual solid particles became more intimate for soils with the relatively higher ρ b in the west pasture.   0-5 cm depth (P < 0.05), while λ value of annual pasture soils was intermediate. Except for alfalfa-TW, the differences in Dt at saturation among native, alfalfa-TW, and OWBlegume followed a similar pattern of λ values for the same soil depth. For the 5-20 cm depth, Dt at saturation did not vary between treatments. At PWP, native had greater Dt than OWB-legume and alfalfa-TW in the 0-5 cm, whereas, in the 5-10 cm depth, OWBlegume and annual had greater Dt than alfalfa-TW (P < 0.05).  (2016-2017). The θv is a function of applied pressure heads given by measured soil water retention curves (SWRCs) (Figure 4). Data pooled from years, depths, and pressure heads applied for determining SWRCs.
Despite having lower θv (a function of pressure heads) in the west pasture soils (Figure 4), the higher λ and Dt in the west compared to the east could be explained by various factors, such as soil texture or particle size distribution, SOM, and ρb [81], which were affected by pasture management practices. Soils with higher clay contents in the east pasture (i.e., clay loam soils) likely possessed a higher degree of aggregation than soils in the west pasture (i.e., sandy clay loam soils) with greater sand contents. This suggested that poorly aggregated soils in the west area had smaller pore spaces with better contact Figure 6. Relationship between soil thermal conductivity (λ) and volumetric water content (θ v ) within the 0-20 cm soil depth for all pasture treatments in the east (solid line) and west (dashed line) area during a 2-year study period (2016-2017). The θ v is a function of applied pressure heads given by measured soil water retention curves (SWRCs) (Figure 4). Data pooled from years, depths, and pressure heads applied for determining SWRCs.
As discussed in Section 3.2, pasture soils with higher total porosity in the east area could decrease soil λ [16,84]. Organic matter, which has been reported to have greater C v and lesser ρ b than mineral soils with a greater pore space [85,86], does not transfer heat as readily as mineral soil. Hence, it was likely that pasture soils with greater SOM in the east area could decrease λ and D t . Our result was in line with previous research (e.g., [20,75], [83]). For instance, Abu-Hamdeh and Reeder (2000) [75] reported a decrease in λ with increasing SOM, i.e., from λ of 0.33 W m −1 K −1 at 5% SOM to 0.17 W m −1 K −1 at 30% SOM. However, Usowicz and Lipiec (2020) [87] observed no clear trend in λ with an increasing application rate of exogenous SOM except for a few measurement dates. SOM content may indirectly affect thermal properties by altering soil aggregate structure [14], in which spherical soil aggregates tend to reduce λ and D t by changing contact areas, and by altering water storage [30], in which soil water mainly increases λ [76].
Based on SWRCs data (Figure 4), the variations in λ and D t values among different treatments were evaluated by three soil water thresholds, namely, very low θ v (at PWP), intermediate θ v (at FC), and high θ v (at saturation). Soil λ and D t at saturation and PWP in different soil depths are presented in Table 4 for the east and west area. Soil λ and D t values in the upper 0-10 cm depth at saturation and PWP were affected by pasture treatments (P < 0.05), whereas λ and D t in the 10-20 cm depth were not affected by treatments (P > 0.05). Soil λ and D t at FC were not affected by pasture treatments (P > 0.05) (data not shown). Soil λ and D t at FC were 1.03 and 1.24 W m −1 K −1 and 0.55 and 0.66 mm 2 s −1 in the east and west areas, respectively, when averaged across years and soil depths. Soil λ and D t values at FC were 1.31, 1.27, 1.37, and 1.31 W m −1 K −1 and 0.54, 0.53, 0.58, and 0.57 mm 2 s −1 for native, OWB-legume, annual, and alfalfa-TW, respectively, in the east pasture area, while in the west area, λ and D t at FC were 1.51, 1.54, 1.55, and 1.60 W m −1 K −1 and 0.65, 0.67, 0.68, and 0.70 mm 2 s −1 for native, OWB-ungrazed, OWB-grazed, and teff, respectively. Neither treatment × depth nor year × treatment interaction was significant for both λ and D t in both pasture areas (P > 0.05). Table 4. Soil thermal conductivity (λ) and thermal diffusivity (D t ) at saturation and permanent wilting point (PWP) soil water thresholds obtained from the soil water retention curves (SWRCs) measured at different soil depths in pasture systems with OWB-legume, native grasses, annuals, and alfalfa-tall wheatgrass for the east and west area during 2016-2017. In the west pasture area, λ had no definitive trend at saturation. For example, native showed greater λ than OWB pastures in the 0-5 cm depth, whereas teff and OWB-grazed had greater λ at saturation than native in the 5-10 cm (P < 0.01) ( Table 4). Soil λ under teff was higher at 0-5 and 5-10 cm depths. Pasture treatments did not affect λ at PWP for all soil depths (P > 0.05). Similarly, D t did not differ among treatments at saturation for all depths (P > 0.05). At PWP, only the upper 0-5 cm depth showed differences in D t , where native had greater D t than OWB-ungrazed (P < 0.01). Teff and OWB-grazed had intermediate D t values between native and OWB-ungrazed. For the 5-20 cm depth, D t did not vary among treatments at PWP (P > 0.05). Overall, greater D t within the upper 0-20 cm soil depths in the west pasture area compared to the east pasture area (Table 4) likely allowed the heat to penetrate deeper into its profile rapidly. Although heat captured at the soil surface needs to be retained to aid forage crop development, greater D t is also desirable to prevent the occurrence of temperature extremes [88], especially in semiarid environments.

Conclusions
The perennial OWB-legume and native pasture systems that contributed to the buildup of SOM and reduced ρ b in the upper 0-20 cm soil depth significantly increased k s compared to annual (i.e., teff and alfalfa-TW) pasture systems (P < 0.05). The perennial pasture systems also improved soil water thresholds (i.e., θ v at saturation, FC, and PWP, as well as PAW) and thermal properties (λ and D t ) than annual pasture systems (P < 0.05), underscoring the benefits of permanent, perennial cover on soil water storage in semiarid environments. Native and OWB-legume pastures, especially in the clay loam east pasture, showed a similar increase in k s and soil water retention characteristics, owing to their greater SOM contents but lower ρ b compared to annual teff and alfalfa-TW (P < 0.05). Soil λ increased with increasing θ v (a function of pressure heads); however, there were no discernible changes in D t with increasing θ v since the increase in C v resulted in very little relative changes in λ and D t of pasture soils across the range of θ v given by SWRCs. Differences in greater λ and D t with differences in soil water thresholds between perennial and annual pastures, especially at saturation and PWP, may be attributed to differences in SOM and ρ b , in addition to a priori differences in soil texture. Grazed pastures decreased k s and soil water retention compared to other treatments (P < 0.05), yet did not affect λ and D t (P > 0.05), likely due to higher soil ρ b and contact areas between particles. Significantly lower soil water retention, k s , λ, and D t in annual and teff pastures compared to perennial systems (P < 0.05) were attributed to annual disturbance, including tillage practices. Overall, our results suggest that management practices involving continued soil disturbance, including annual forage crops and tillage-based row-crop agriculture, are less beneficial to water storage and heat movement in unsaturated soils than perennial pasture systems, especially in semiarid regions.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12071491/s1: Table S1: Description of forage and pasture management practices under different treatments over the years in the east pasture area at New Deal, TX, USA. Pasture treatments include native, Old-World Bluestem (OWB) (OWB, Bothriochloa bladhii (Retz) T. Blake)-legume, annual, and alfalfa-tall wheatgrass (TW) (TW, Thinopyrum ponticum (Host) Beauv.). Area is for a single replicate. Table S2: Description of forage and pasture management practices under different treatments over the years in the west pasture area at New Deal, TX, USA. Pasture treatments include native, teff, Old World bluestem (OWB) (OWB, Bothriochloa bladhii (Retz) T. Blake)-ungrazed, and OWB-grazed. Area is for a single replicate.