Influence of Rangeland Land Cover on Infiltration Rates, Field-Saturated Hydraulic Conductivity, and Soil Water Repellency in Southern Patagonia

ABSTRACT This study investigated the influence of rangeland land cover on infiltration rates (IRs), field-saturated hydraulic conductivity (Kfs), and soil water repellency in Patagonia. Four land cover types (shrubs, dwarf heath shrubs, bare soil, and Inter tussock) were examined to assess their effects on hydrological processes. IR was measured using the single-ring method, and soil water repellency was evaluated using the Beerkan method. We hypothesized that land cover type affects IR, Kfs, and soil water repellency. The results showed significant variations in IRs among land covers, with Tsf displaying lower rates than the other covers. Soil water repellency was prevalent in shrub and dwarf heath shrub–covered soils. Lateral flow was observed, indicating limited water infiltration. The obtained Kfs values were higher than the calculated hydraulic conductivity values (Ks). However, further investigation is required to assess the impact of capillarity (i.e., α*) on Kfs determination. This study enhances our understanding of hydrological processes in rangeland ecosystems and provides valuable insight into land management practices. By elucidating the relationships among land cover, IR, Kfs, and soil water repellency, this study contributes to sustainable water resource management in arid and semiarid regions.


Introduction
Understanding soil surface water infiltration rates and saturated hydraulic conductivity is essential for managing soil and groundwater recharge, mitigating drought, promoting soil health, and evaluating soil improvement strategies. This knowledge is fundamental in the context of climate change, which poses significant challenges to soil water management ( Rozanski 1985 ;Shukla 2014 ;Concialdi et al. 2020 ).
Water infiltration into soils is affected by various factors, including soil texture, organic matter content, and soil water repellency or hydrophobicity. The latter refers to reducing the soil affinity to water, which can resist wetting for varying periods, ranging Saturated hydraulic conductivity (K s ) is also essential for understanding water storage and recharge in soils, groundwater recharge, surface runoff, and erosion. Although its measurement is sometimes complex and requires specialized laboratory equipment, the estimation of field-saturated hydraulic conductivity (K fs ) allows for multiple measurements in a brief time and is within reach of small-budget projects and time ( Lassabatère et al. 2006 ;Bagarello et al. 2014 ).
Rangelands cover approximately 54% of the terrestrial area of the world, sustaining 69% of all agricultural land (e.g., livestock, meat, and dairy, including grazing land and arable land fields) used for animal growth and feed production ( Bardgett et al. 2021 ;Magliano et al. 2023 ). Therefore, rangelands are essential for sustaining productivity, economy, and livelihoods ( Krueger et al. 2021 ). One of the essential variables to explain rangeland productivity is soil water content, which explains 60% of ecosystem productivity ( Recous et al. 2008 ;Krueger et al. 2021 ). However, rangelands comprise several shrub and grass species in different phenological states, usually forming a unique ensemble. The presence or absence of some species affects soil bulk density, root distribution, root thickness, soil organic matter content, and microbial activity, all parameters known to affect water infiltration in soils differently ( Doerr et al. 20  The southernmost region in Chile, Magallanes, is mainly covered by rangelands. The region holds 91% of rangelands to sustain livestock in Chile ( Radic-Schilling et al. 2021 ). However, these lands have been subjected to approximately 150 yr of extensive livestock grazing, primarily sheep Radic-Schilling et al. (2022) . In addition, the native niches of grass and shrub species showed different degrees of degradation. Common meteorological conditions in Magallanes include high wind speeds throughout the yr (annual average of 10 m s −1 ), with precipitation ranging from 100 to 320 mm yr −1 . In addition, low temperatures and high radiation occur during the austral spring and summer ( Radic-Schilling et al. 2021 ;Butorovic 2016 ). Therefore, in this region, ecosystem services, such as water infiltration, water storage in soils, and hydrophobicity, have become increasingly important, particularly under the current climate change drought scenario over the past decade. However, despite this information, little to no information is available on the variables mentioned earlier and how these variables are affected by vegetation. Therefore, the working hypothesis of this study is as follows: The water infiltration rate (IR), field-saturated hydraulic conductivity (K fs ), and soil hydrophobicity vary significantly among different land cover types (shrubs, dwarf shrub heath, and inter tussock [IT]) in the Magallanic rangelands. We expect to observe differences in these soil properties based on each land cover type's characteristics and ecological functions.

Study site
Our study site is 80 km from Punta Arenas, in "Estancia Josefina." The climate at the study site is classified as steppe or cold semiarid (BSk, according to Köppen-Geiger; see Fig. 1 ) ( Radic-Schilling et al. 2021 ). On average, the study site precipitation was approximately 320 mm yr −1 (2013-2021 period (INIA, Kampenaike). However, the annual rainfall in 2022 is 182 mm. Infiltration measurements were performed under the main land cover types as follows (see Fig. 1 ): (1) Shrubs (Sh), mainly dominated by ( Chiliotrichum diffusum [g. forst.] Kuntze), with the presence of other species such as Lepidophyllum cupressiforme , Mulgurea tridens , and Adesmia boronoides y Mulinum spinosum . However, all measurements were performed under Ch. diffusum ( Fig. 2 A ). Its height ranges from 0.3 m to 1.5 m.
(2) Dwarf heath shrub (Dsh) Empetrum rubrum Vahl dom . Ex Willd. (Red Crinatesowberry). However, other species like Baccharis magellanica , Berberis empetrifolia , Azorella caespitosa , Nardophyllum bryoides , Bolax gummifera, and Azorella trifurcate are present in a minor proportion. This last land cover is found in flat to low-slope terraces, exposed to wind, and thin soils subjected to wind erosion, with pebbles on the surface but smaller soil particles below the latter ( Radic-Schilling et al. 2021 ). Dwarf heath shrub land cover height is around 0.05 m to 0.1 m and grows to form patches surrounded by bare soil, as described earlier. Therefore, bare soil (Bs) was selected as the land cover type (see Fig. 2

B). (3) Inter Tussock (IT) or the area between Festuca gracillima
Rothm . Individuals. F. gracillima is the dominant species in the region and usually covers 30-70% of the area ( Radic-Schilling et al. 2021 ). Its height ranges from 0.1 m to 0.5 m in the area where the measurements occurred. IT comprises a wide range of native graminoid species (see Fig. 2C ).
Infiltration rates and field-saturated hydraulic conductivity measurements: Infiltration rates (IR, in mm h −1 ) and K fs were measured and calculated for each land cover type using the Beerkan methodology (described in Lassabatère et al. 2006 ). Under each land cover, infiltration measurements were taken in a single 4-m transect with nine infiltration rings separated by 0.5 m. The infiltration ring had a 10.7-cm inner diameter of 4.7 cm high. In the Beerkan method, vegetation is removed from the surface, leaving only the roots intact. A single ring was then inserted to a depth of approximately 1 cm to prevent lateral loss of ponded water from the soil surface. Please note that the Beerkan method measures infiltration in three dimensions (sideways and downward), unlike a double-ring infiltrometer, which measures water infiltration in soils in one dimension (only downward). Subsequently, a fixed volume of water (in this study, 100 mL, representing approximately 9 mm) was poured into the cylinder at time zero, and the time elapsed during the infiltration of the known volume of water was measured. When the first volume had thoroughly infiltrated, a second known volume of water was added to the cylinder, and the time required to infiltrate was measured (cumulative time). This procedure was repeated until three steady measurements were recorded ( Angulo-Jaramillo et al. 2016 ). The obtained data were used to calculate the initial infiltration rates (IIR, in mm h −1 ) at time = 0, using nonlinear regression fitting, infiltration rates (IR, in mm h −1 ) during measurement, and accumulated infiltration (AI, in mm), corresponding to the total accumulated volume at a specific period during the infiltration measurement. The field-saturated hydraulic conductivity (K fs , in mm h −1 ) was calculated as follows ( Angulo-Jaramillo et al. 2016 ): Where "b 1 " and "r" stand for infiltration slope (in mm h −1 ) and ring radius (in mm), respectively. While α * stands for: Where ϕ m (mm 2 h −1 ) is the matrix flux potential ( Elrick and Reynolds 1992 ). According to the literature, the α * parameter can be estimated on the basis of a general description of the textural and structural characteristics of the soil ( Bagarello et al. 2014 ).
Because soils under each land cover were classified as sandy loam soils, the α * value used in this study was 0.012 mm −1 . This value is  also suggested by Angulo-Jaramillo et al. (2016) and Bagarello et al. (2014) , corresponding to "most soils with a structure and medium and fine sands" ( Elrick and Reynolds 1992 ) or with "Moderate" capillarity (see Bagarello et al. (2014) for details on the derivation of Eq. 1 ). Before starting an infiltration run, the top 10 cm of soil was sampled (within a distance of 0.3 m) to determine its initial gravimetric water content. Once the infiltration experiments were completed, saturated soil was sampled to determine the saturated gravimetric water content. Water content was determined by drying the samples at 105 °C until a constant weight was achieved. All samples were transported in independent, labeled, and closed plastic bags. The initial gravimetric content is subtracted from the saturated gravimetric content. This was recorded as " Soil Moisture Content (in %)". These values were used to determine the amount of water gained from the soil during infiltration runs.
Antecedent precipitation, soil description, water content, texture, and organic matter content: Antecedent precipitation (AP) was calculated as the sum of precipitation for 1 d, 7 d, and 14 d before the infiltration measurement. The antecedent potential evapotranspiration (AETo) was calculated using the Penman-Monteith method for the same period.
The measured AP and AETo in the previous 14 d, 7 d, and 1 d were 10.9 mm, 6.5 mm, and 0 mm, respectively, with 14 d, 7 d, and 1 d of potential evapotranspiration and 5.7 mm, 2.9 mm, and 0 mm, respectively. Annual ETo at the site was approximately 700 mm yr −1 , and a similar value was determined by Radic-Schilling et al. (2021) .
Under the selected land covers, soils were sampled and analyzed for initial water content, soil texture, and organic matter content. The soil water content was determined by weight loss after drying at 105 °C. After this procedure, each sample's subsample (4-5 g of soil) was sieved through a 2-mm mesh. The < 2 mm fraction was used to calculate the organic matter content through the weight loss after ignition (360 °C for 16 h). Soil texture was determined using 125 μm (very coarse sand to fine sand), 32 μm (very fine sand and coarse silt), and smaller particles (medium silt to clay particles).
Because the obtained data were not normally distributed or homoscedastic, they were compared using a nonparametric Kruskal-Wallis (K-W) test with a posteriori comparison of Dunn's test in case statistical differences were found. The statistical significance level was set to 0.05 ( α = 0.05). All presented results are presented as an average ± 1 standard error. Table 1 Land covers, initial and final soil water content (SWC, mean ± 1 standard deviation, in %), texture (in %) and organic matter content (OM, in g) for soils under each land cover used in this study. Statistical grouping is shown with numbers 1 to 3 ( P < 0.05; Dunn). N = 9; 8; 5 and 12 samples for shrubs (Sh), dwarf shrub heath (Dsh), barren soil (Bs), and inter tussock (IT), respectively.
Textural analysis showed that all soils under different land cover types were classified as sandy loam (see Table 1 ). Generally, the sand content was statistically lower in IT than in Sh, Dsh, or Bs (see Table 1 ). Consequently, finer soil particles were statistically higher in IT than in Sh, Dsh, and Bs (see Table 1 ). Dsh and Bs showed no statistical differences in soil texture and organic matter content ( P > 0.5, K-W). Despite this, Dsh and Bs were considered different because root channels might affect the conducting of infiltrated water faster in the soil under Dsh and nearby Bs.

Infiltration rates and accumulated infiltration for each land cover type
The infiltration rates (IR, in mm h −1 ) are shown in Figure 3 . IR showed statistically significant differences ( P < 0.05, K-W). However, only IT was statistically different from Sh, Dsh, and Bs ( P < 0.0 0 01 for all land cover types, Dunn's test). However, Sh, Dsh, and Bs showed no statistically significant differences ( P > 0.2, Dunn's test). Dsh and Bs showed no statistically significant differences ( P < 0.9, Dunn's test) (see Fig. 3 ).
The IR and Accumulated Infiltration (AI, in mm) are shown in Figure 4 (right panel). IR showed high variability (Min-Max and range) in soils under Sh (13.6-699.7, range: 686.1 mm ·h −1 ); Dsh (7.7-729.8, range: 722.1 mm ·h −1 ); Bs (3.7-4 85.8, range: 4 82.0 mm ·h −1 ). IT showed the lowest IR variability (3.6-307.2, range: 303.6 mm ·h −1 . Consequently, the IR in soils under the studied land cover types can be ordered as follows: Dsh > Sh > Bs > IT. AI Min-Max and range observed values were the following: Sh (71.7-241.9, range: 170.2 mm), Dsh (116.5-241.9, range: 125.4 mm), Bs (71.7-197.1, range: 125.4 mm), and IT (80.6-179.2, range: 98.6 mm). Therefore, according to their AI values, the soils under the studied land cover could be ordered as follows: Sh > Dsh = Bs > IT. However, this variable showed no statistical differences when comparing land cover ( P > 0.6, K-W). IR under Sh and Dsh showed a steep decrease during the first minutes of the infiltration runs, followed by a slow but steady increase in infiltration rates (see Fig. 4 , left panel for Sh and Dsh). While AI was higher under Sh and Dsh, it was considerably lower under the IT cover (see Fig. 4 ).

Variation in soil water content after the infiltration runs
The initial soil water content (SWC, in %) under the selected land cover showed statistical differences ( P < 0.05; K-W). The statistical groupings were as follows: Sh and IT and Dsh and Bs in the other group (see Table 1 ). After the infiltration runs, SWC only differentiated Sh from the other three land cover types ( P < 0.05; Dunn). This study used the difference between the final and initial SWC as a proxy for soil water repellency. Our results clearly show that Sh and Dsh had the highest frequencies (6/9 and 2/9, for Sh and Dsh, respectively) of infiltration runs that presented water repellency (see Fig. 4 ). However, Bs and IT exhibited the lowest frequencies (1/5 and 0/9, respectively).
The infiltration runs under Sh showed a slight difference between the final and initial soil water content (see Table 1 and Fig. 4 ). Although the rings were properly inserted during the infiltration runs under Sh, a lateral surface flow was observed. All infiltration measurements that showed water repellency showed a final SWC value of < 30% (see Table 1 and Fig. 5 ).

Field-saturated (K fs ) and modeled (Kfs model), and K s hydraulic conductivity
K fs showed no statistical differences when this variable was compared for each land cover type ( P = 0.3; K-W, see Fig. 6 ). The same trend was observed for K s . K fs under each land cover type could be ranked as follows: Dsh > Sh ≈ Bs > IT (see details in Table 2 ). Similar results were observed for the Kfs-modeled and K s (see Table 2 and Fig. 6 ).   [Bs], and inter tussock [IT]), initial infiltration rate (IIR, in mm h −1 ), infiltration rate (IR, in mm h −1 ), field-saturated hydraulic conductivity (K fs , in mm h −1 ) calculated according to Bagarello et al. (2014) using the last three measurements; K fs model, as the value of the asymptote of the nonlinear regression (K fs model, in mm h −1 ), hydraulic saturated conductivity (K s in mm h −1 ) calculated as in Bagarello et al. (2014) , and the number of infiltration runs used for the calculation of previously mentioned variables (N). 423.6 ± 17.9 16.9 ± 0.7 7.4 ± 4.6 10.7 ± 0.7 2.9 ± 1.8 12 Figure 5. This plot shows the variation of soil water content (as a percentage) for each of the measurements taken under different soil land covers, calculated as the final soil water content minus the initial soil water content. The black dashed line indicates a 7% variation in soil water content. Note that any bar lower than the dashed line indicates evidence of soil water repellency.

Infiltration rates
Our study found that IR varies depending on the rangeland land cover type. However, this difference may be attributed to the higher organic matter content in IT soil and finer particles in Sh and IT soils. In this sense, IT showed lower values for IR and IIR than those observed for Sh, Dsh, and Bs (see Table 2 ). It is known that larger soil particles highly influence the IR and IIR. This was clearly the case for the IT observed difference with the IR and IIR values from Sh, Dsh, and Bs (see Table 2 ). This could be related to the finer textured soil particles in IT, which had almost twice the amount of finer-texture particles compared with those of Sh, Dsh, and Bs. It is widely known that soil texture is an important variable that affects water infiltration in soils. Our study classified the soils under each land cover type as "sandy loam." Sh, Dsh, and Bs show similar IR values (see Fig. 3 ). These infiltration rates are low but within the range of those described in the literature for land cover type ( Shukla 2014 ;Fraser and Stone 2016 ) and those described for similar land covers in northern Patagonia ( Chartier et al. 2011 ). The lower infiltration rates observed in IT may be related to a thinner, grouped, and downward radicular system and a higher proportion of finer soil particles ( Braud et al. 2005 ;Shukla 2014 ;Fraser and Stone 2016 ;Lichner et al. 2018 ;Di Primma et al. 2021 ).
Although the initial conditions, including soil texture and water content, were similar for Sh and IT, they exhibited different IR and IIR values. We attribute these differences to soil texture. Soil texture, especially sand content, significantly affects the IR and IIR ( Braud et al. 2005 ;Shukla 2014 ;Wu et al. 2022 ). In contrast, Dsh and Bs showed similar initial SWC, texture, and organic matter contents. Therefore, it may be easier to compare them. However, this was not the case because IR and IIR are also distinct. Root density was not measured but could explain IR and IIR. Sh has a thick understory of grass and mosses. Although the understory of IT was visually less dense, there were many grass species. It is known that roots from large (e.g., from trees and shrubs) and small (e.g., from mosses and grasses) plants play an essential role in conducting water through the surface soil ( McGlynn et al. 2002 ;Scherrer and Naef 2003 ;Lichner et al. 2007b ;Mohammed et al. 2018 ).

Soil water repellency
Soil water repellency is common in soils with high organic matter/compound content ( Hallett 2007 ). Frequently, these soils are subjected to constant drying ( Doerr et al. 20 0 0 ;Hallett 20 07 ) or occur after fires ( Ferreira et al. 20 0 0 ;Garcia-Chevesich et al. 2010 ;Oyarzún et al. 2011 ) or by the presence of microorganisms ( Hallett et al. 2001 ;Hallett 2008 ). Patagonian rangelands thrive over young and shallow organic soils and are subject to constant winds during the growing season (Austral spring-summer seasons). Therefore, a certain degree of water repellency is expected. The Sh and Dsh land cover types exhibited the highest frequency of infiltration runs, indicating soils with water repellency (66% and 22%, respectively). The measurements made under shrub land cover initially showed high infiltration rates. However, lateral flow was observed in 67% of the infiltration runs. There was no visual evidence of any physical factor (e.g., rock, pebbles, or > 2 mm roots) that could explain the lateral flow or the lack of water in the middle or bottom sections of the 10-cm soil profile. Some authors have reported that organic matter content is frequently linked to water repellency ( Huber et al. 2010 ;Abel et al. 2013 ;Concialdi et al. 2020 ) but also to dry soil, organic compounds, and microorganisms ( DeBano 20 0 0 ; Doerr et al. 20 0 0 ;Doerr and Thomas 20 0 0 ;Garcia-Chevesich et al. 2010 ;Hallett et al. 2001 ;Lichner et al. 2007a ;2008 ). In our study, the organic matter content in soil samples under Sh land cover was statistically different from that of Dsh, Bs, and IT. This distinction could explain the differences in IR behavior and the high soil hydrophobicity observed during infiltration runs in soils under Sh land cover. However, a more detailed study is required to investigate whether the observed hydrophobicity is related to a specific organic compound ( Gregory et al. 2005 ;Lichner et al. 2018 ).
The 30% initial soil moisture content boundary suggests that soils below that threshold exhibit water-repellent qualities, as described by Bayad et al. (2020) , Clothier et al. (20 0 0) , andLichner et al. (2007a) . Further exploration of other variables, such as the polarity of organic compounds reaching the soil surface or generated as by-products of organic matter decomposition and temperature interaction, should be considered in future studies to characterize the repellency and promote further soil water recharge ( Hallett 2007 ;Walden et al. 2015 ;Lichner et al. 2018 ).
Water filling on the outer sides of the infiltration ring was observed in the field. Therefore, water infiltrated the top layer of the soil, but the lateral flow was observed in all directions. A slightly concave shape characterizes the accumulated Infiltration under Sh and Dsh, followed successively by a linear part and a final somewhat convex line (see Fig. 4 ). The literature describes these soils as typical water-repellent soils ( Clothier et al. 20 0 0 ;Alagna et al. 2016 ). Water infiltration was fast in the beginning. However, it was impeded later by some of the following factors: 1) soil is too dry and/or 2) high organic matter/compound content ( Clothier et al. 20 0 0 ). It is widely recognized that the presence of organic matter and certain organic compounds can diminish water infiltration in soils ( Doerr et al. 20 0 0 ;Oyarzún et al. 2011 ;Abel et al. 2013 ;Di Primma et al. 2021 , among others). Organic substances can hinder soil water recharge, a critical process in dry environments such as the Patagonian rangeland ( Cavallaro et al. 2020 ;Radic-Schilling et al., 2021 ). "K fs ," "K fs model," and "K s " In our study, the obtained K fs values were 2.6 × higher than those of K s obtained using the methods described earlier. "K fs model" values, calculated using nonlinear regression fitting, resulted in higher "K fs model" values in soils under Sh, Dsh, and Bs. For IT, the values were lower than those of the other land covers measured in this study. The observed K fs and K s values were within the values reported for sandy loam soils in the literature ( Carsel and Parrish 1988 ;Shukla 2014 ;Islam et al. 2019 ). The K s values were lower than those of K fs . We attribute this difference to the fact that Ks represents a unidimensional measurement (i.e., downward flux), whereas K fs represents a three-dimensional measurement (i.e., downward and sideways) of hydraulic conductivity. ( Elrick and Reynolds 1992 ). The three-dimensional measurement of hydraulic conductivity, K fs , is more complex because the waterfront within a single measurement grows exponentially, whereas using unidimensional methods, the waterfront is more stable and continuous ( Elrick and Reynolds 1992 ;Braud et al. 2005 ). If soil heterogeneity is included, K fs will have higher values than K s . According to Bagarello et al. (2014) , an incorrect choice of α * by category (i.e., 0.0 01 mm −1 , 0.0 04 mm −1 , 0.012 mm −1 , and 0.036 mm −1 ) should not substantially compromise the reliability of the K fs prediction for most field soils. In the present study, the parameter α * (i.e., capillarity) was not measured; therefore, it cannot be discarded as a source of error in the determination of K fs .
Single-ring infiltration measurements and K fs determination using the Beerkan method are field friendly (e.g., simple, use less water) and less prone to error than other infiltration measurement methodologies (e.g., the double-ring infiltrometer method) or two or more ponded heads in an uncased well (e.g., The Guelphpermeameter method), which is much more susceptible to errors induced by local small-scale heterogeneity because the infiltration surface changes with each ring radius or ponded head ( Reynolds and Elrick 1990 ;Bagarello and Lovino 2012 ;Lassabatere et al. 2019 ).

Conclusion
This study reports the first use of the Beerkan method to measure infiltration rates, K fs , K s , and hydrophobicity under common land cover types (Shrubs, Dwarf shrub heath, bare soil, and Inter tussock) in Patagonian rangelands. Only Inter tussock showed significantly different measured infiltration rates than the other three land cover types. During the infiltration runs, high water repellency was observed under Shrubs and Dwarf shrub heath (67% and 25%, respectively). In contrast, IT and bare soil showed the lowest water repellency (8% and 0%, respectively). K fs values showed high variability and were not differentiated among the land cover types. However, K fs values are twice as high as the calculated K s values. Therefore, the K fs and K s values calculated in this study should be used with caution.
It is important to acknowledge the limitations of this study, which was focused on a single site. Future research should encompass a wider range of sites to capture a more comprehensive understanding of the variations in infiltration rates, K fs , and soil water repellency. Additionally, including more soil analyses, such as soil structure, pore size distribution, and organic compounds that promote water repellency would enhance insights into the underlying mechanisms and interactions affecting these variables. In addition, further research should be focused on studying the relationship between hydrophobicity and soil health status, as well as their implications for soil and water management in dry environments such as the Patagonian rangeland.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Author contributions
P. H-F and S. R. designed the experiment, execution and wrote the final version of the manuscript; P. H-F, R. M., C. G., and L. B. performed field and laboratory measurements. P. H-F. wrote the first draft. S. R. Funding Acquisition. All authors contributed equally to the discussion regarding further improvement of the manuscript.