The study was conducted on a private farm in Desha County, southeast Arkansas within Major Land Resource Area (MLRA) 131B, Arkansas River Alluvium, which is an alluvial plain composed of variable soil textures, bottom-land hardwood deciduous forests, and cypress (Taxodium distichum) swamps [13][18]. The specific study area was located approximately 46 m downslope (<1%) from the northern edge into production-scale field and was approximately 49 m long and 11.4 m wide with 12, 0.7-m wide raised beds [13]. The soil mapped throughout the study area was a Hebert silt loam (fine-silty, mixed, active, thermic Aeric Epiaqualfs), which is a poorly drained and slowly permeable soil developing in silty alluvium [18].

The 30-year (i.e., 1991 to 2020) average annual precipitation in the region was 130 cm, with the greatest monthly rainfall generally occurring in April (i.e., 14 cm) [13][19]. The 30-year monthly average air temperature is 17.8˚C, with the largest average monthly air temperature generally occurring in July (i.e., 27.7˚C) [13][19].

The study area was comprised of four blocks (i.e., replications) with three annual biochar rate treatments (i.e., 0, 2000, and 4000 kg ha⁻¹ yr⁻¹) randomly assigned per block [13][17]. Total biochar amounts applied were 0, 6,000, and 12,000 kg ha⁻¹ total. Consequently, the experimental design was a randomized complete block (RCB) with a factorial design to assess potential effects of biochar rate on soil water retention [13][17]. The biochar treatment rates corresponded to 0, 7.75, and 15.5 kg of surface-applied biochar per plot, and biochar was broadcast-applied to the soil surface one week before planting in plots that were 10 m long and four raised beds wide (i.e., 3.8 m), which represented the experimental unit [13][17] The biochar was derived from a softwood pine (Pinus spp.) feedstock (EnviraPAC, Monticello, AR) and was composed of a mixture of coarse and powder-like particles [13][17]. The biochar material applied had 93% organic C and 0.7% total N [20]. During planting, the applied biochar was lightly mixed in the upper ~5 cm of soil by the mechanical action of the planter [13][17]. The minimal incorporation was preferred by the landowner to preserve the integrity of the raised beds for future use and to allow unobstructed water flow through the adjacent furrows [13][17].

The field study initially began in Spring 2023 with the first of three consecutive years of consistent biochar treatment (i.e., 0, 2000, and 4000 kg ha⁻¹), where biochar was applied to the same plots in Spring 2024 and 2025 as well [13][17]. Burke et al. [17] summarized selected initial biochar properties. The study area has a history of minimum tillage and furrow-irrigation each year when needed with nearby well water (i.e., groundwater) [13][17]. The typical annual tillage performed was one pass of a roller bedder to form or re-form the raised beds and a subsequent pass with a bed conditioner to level the tops of the beds for planting [13][17].

During the 2023 and 2024 growing seasons, cotton was planted on top of the raised beds in April at a rate of ~168,000 seeds ha⁻¹ to a depth of ~2.5 cm, with an in-row spacing of ~15 cm; the raised-bed width was 65 to 70 cm and the bed height was ~9 cm [13][17]. In both growing seasons, ~120 to 130 kg N ha⁻¹ were applied as a combination of either 32% urea ammonium nitrate (UAN), diammonium phosphate (DAP), or ammonium sulfate after planting [13][17]. In addition, in both growing season, 44 kg ha⁻¹ of phosphorus (P) were applied as DAP, 101 kg ha⁻¹ of potassium (K) were applied as muriate of potash, 24 kg ha⁻¹ of sulfur (S) were applied as ammonium sulfate, and 1.1 kg ha⁻¹ of boron (B) were applied as elemental B [13][17].

During the 2025 growing season, corn was planted in late March at a rate of ~84,000 seeds ha⁻¹ to a depth of ~5cm, with an in-row spacing of ~6.4 cm [13][17]. Immediately after planting in 2025, 112 kg ha⁻¹ of ammonium sulfate, 196 kg ha⁻¹ of DAP, 140 kg ha⁻¹ of muriate of potash, and 1 kg ha⁻¹ of chelated zinc were broadcast-applied [13][17]. In late April 2025, 252 kg ha⁻¹ of urea were broadcast, followed by a final urea broadcast application of 112 kg ha⁻¹ in early June, and pesticides were applied each year on an as-needed basis to control weeds [13][17]. Fertilizers and pesticides were applied with a ground rig or via airplane when necessary. Additional field management details were reported in Brye [13] and Burke et al. [17].

At the end of the 2025 corn growing season, one soil sample was collected with a 4.8-cm-diameter, stainless steel core chamber and slide hammer from the top 15 cm of one raised bed in each plot per treatment (n = 4 per treatment) for BD determination and chemical analyses [13]. Soil samples were oven-dried for at least 48 hours at 70˚C, ground, and sieved though a 2-mm mesh screen [13].

Soil pH and electrical conductivity (EC) were determined potentiometrically in a 1:2 soil mass:water volume suspension [21]. Plant-available Ca, Mg, K, and P concentrations were determined following Mehlich-3 extraction in a 1:10 soil mass:extractant volume and analysis by inductively coupled, argon-plasma spectrophotometry [22]. SOM concentration was measured after 2 hours of combustion at 360˚C by weight-loss on-ignition [23]. TC and total N (TN) concentrations were measured on a VarioMax CN analyzer (Elementar Americas Inc., Mt. Laurel, NJ) by high-temperature combustion [24][25]. All measured TC was assumed to be organic due to no effervescence when treated with dilute hydrochloric acid. The soil C:N ratio was calculated from the measured TC and TN concentrations.

A second soil sample was collected with a 5-cm-tall core chamber, which was manually tapped into the soil with a rubber mallet, from the top 5 cm of one raised bed in each plot per treatment (n = 4 per treatment) for water-retention measurements. Samples were air-dried for at least 5 days at ~21˚C, then ground and sieved with light pressure though a 2-mm mesh screen.

Following prior procedures [4][26]-[28], 10, 5.0 ± 0.1 g subsamples of air-dried, ground, and sieved soil from each of the 12 plots were placed into small, plastic cups, after which 1, 2, 4, 6, 10, 12, 15, 20, 30, and 40 drops of distilled water were added, respectively, to achieve a representative range of gravimetric soil water contents. Any dispersive effects from using distilled water were assumed to affect all soil samples equally. Following the addition of distilled water, the soil samples were manually mixed with a metal spatula to uniformity and subsequently transferred to pre-weighed, small, plastic measurement capsules (i.e., 4.0 cm in diameter and 1.1 cm tall; Decagon Devices, Inc., Pullman, WA). Soil was placed to fill approximately half the height of the measurement capsule, after which the capsule was capped and allowed to equilibrate overnight at room temperature (i.e., ~21˚C). The following morning, the soil water potential of each sample was then measured with a Decagon WP4C Dewpoint Potentiameter (Decagon Devices Inc.). After each water potential measurement, the capsule, cap, and moist soil were weighed and oven-dried at 70˚C for at least 12 hours. After oven-drying, the capsule, cap, and dry soil were reweighed to calculate the gravimetric water content of each sample based on the initial air-dried mass of soil. The resulting water potential measurement associated with each gravimetric soil water content was used to represent the soil wetting curve for all replicate samples.

The relationship between soil water potential (i.e., dependent variable) and gravimetric soil water content (independent variable) was characterized via linear regression using Excel (version 2607, Microsoft Corp., Redmond, WA) by fitting trendlines through all data points for each replication of the three biochar treatments (n = 4) [26][28]. A simple linear regression equation was determined for both the raw (i.e., non-transformed) and natural-logarithm (LN)-transformed gravimetric soil water content data points. The linear and LN slopes and intercepts of the resulting regression relationships were recorded for each biochar plot for subsequent statistical analyses [26][28].

Based on the RCB design, a one-factor analysis of variance (ANOVA) was performed in SAS (version 9.4, SAS Institute, Inc., Cary, NC) to determine the effect of biochar treatment (i.e., 0, 2000, and 4000 kg ha⁻¹) on soil BD, pH, EC, extractable soil nutrient (i.e., Ca, Mg, K, and P) concentrations, SOM, TC, and TN concentrations, and the soil C:N ratio in the top 15 cm after three consecutive years of consistent management. All soil properties were analyzed with a gamma distribution. A one-factor ANOVA was also performed in SAS, based on a RCB design, to determine the effect of biochar treatment on the linear and LN slope and intercept parameters from the regression fits characterizing the relationship between the measured soil water potential and gravimetric soil water content from the top 5 cm. The slope parameters were analyzed with a gamma distribution due to the skewed data distribution, while the intercept parameters were analyzed with a normal distribution.

Treating the biochar rates as a continuous variable, rather than a discrete variable as with the ANOVA analyses, linear regression analyses were also performed on all soil properties and the slope and intercept parameters from the linear and LN fits characterizing the relationship between the measured soil water potential and gravimetric soil water content (Microsoft Excel 16). Significance was judged at P ≤ 0.05 for all analyses.

After the three consecutive years of pre-plant biochar additions (i.e., 2023, 2024, and 2025), soil BD, pH, EC, extractable Ca, Mg, K, and P, and SOM, TC, and TN concentrations and C:N ratio in the top 15 cm did not differ (P> 0.05) among biochar treatments when biochar treatments were treated as a discrete or continuous variable (Table 1). However, soft, woody feedstocks, which served as the main component of the biochar used in this study, and softwood-derived biochar have been shown to increase soil K levels [29]. In addition, during pyrolysis, K is mostly preserved and converted into highly soluble salts [30] that are easily released upon biochar interaction with soil and soil moisture fluctuations. Potassium is also often one of the most abundant plant macronutrients in many soils [31], which may have masked biochar-addition effects on soil K.

Table 1. Summary of the effect of biochar (B) treatment [i.e., 0 (0B), 2000 (2B), and 4000 (4B) kg ha⁻¹] on bulk density (BD), soil pH, electrical conductivity (EC), Mehlich-3 extractable soil nutrient (i.e., Ca, Mg, K, and P) concentrations, soil organic matter (SOM), total carbon (TC), and total nitrogen (TN) concentrations, and the soil C:N ratio (C:N ratio) in the top 15 cm after three consecutive years (i.e., 2023, 2024, and 2025) of biochar addition (n = 4 per treatment) in a furrow-irrigated, cotton-corn rotation in southeast Arkansas.

Though biochar is known to be a generally recalcitrant C source [9], the first-year applied biochar (i.e., 2023), which also resulted in twice the biochar-C application in the high- compared to the low-biochar-rate treatment, likely had started to decompose by year three (i.e., 2025) to stimulate microbial activity [32], specifically as carbon dioxide (CO₂) emissions [33], to contribute to numerically, though not significantly, decreasing TC concentrations as biochar rate increased (Table 1). Brye [13] reported season-long CO₂ emissions numerically increased as biochar rate increased, but season-long CO₂ emissions differences were not significant among biochar rates (i.e., 0, 2,000, and 4,000 kg ha⁻¹). The numerically, though not significantly, decreasing soil C:N ratio after three consecutive years of biochar application was a positive influence on soil health by creating a more conducive soil environment for microbial activity [34][35] and may have been reflective of the decomposition of the first year of biochar addition after three cropping season years. The lack of a significant effect of biochar rate on soil properties reflected larger variability among individual treatment replications and were likely related to the minimal biochar incorporation that occurred each application year [13].

Combining data across all three biochar treatments and four replications, the relationship between measured water potential and gravimetric water content in disturbed fine-earth material was curvilinear, where the water potential increased exponentially as gravimetric water content increased (Figure 1). Similar patterns have been reported in other studies on loamy soils in eastern Arkansas portion of the LMRV [4][27][28].

Figure 1.Relationship between soil water potential and gravimetric soil water content for three biochar treatments and all replications within a treatment (n = 4) for the top 5 cm after three consecutive years (i.e., 2023, 2024, and 2025) of biochar addition in a furrow-irrigated, cotton-corn rotation in southeast Arkansas.

Similar to most end-of-season soil properties, but contrary to the original hypothesis, whether treated as a discrete or continuous variable, the slope and intercept parameters characterizing the both the linear and LN relationship between the measured soil water potential and the gravimetric soil water content did not differ (P> 0.05) among biochar treatments after three years of biochar application and three cropping seasons (Table 2). The predictive ability of the linear trendlines was not large, where coefficients of determination (r²) values ranged from 0.21 to 0.37, while the r²values were somewhat larger for the LN relationships, which ranged from 0.49 to 0.79 across individual replications.

Table 2. Analysis of variance summary of the effect of biochar (B) treatment [i.e., 0 (0B), 2000 (2B), and 4000 (4B) kg ha⁻¹], treated as a discrete variable, on the slope and intercept parameters from the linear and natural logarithmic (LN) fits characterizing the relationship between the measured soil water potential and gravimetric soil water content and the slope and intercept parameters from linear regression analysis treating biochar rates as a continuous variable in the top 5 cm after three consecutive years (i.e., 2023, 2024, and 2025) of biochar addition (n = 4 per treatment) in a furrow-irrigated, cotton-corn rotation in southeast Arkansas.

Soil water retention, as characterized by soil wetting curves, was unaffected by three consecutive years of biochar treatment to an alluvial, silt-loam soil in southeast Arkansas. Results of the current study were consistent with Gaskin et al. [36] who concluded that high rates of biochar application (i.e., 22 Mg ha⁻¹) are needed in order to increase soil water retention in loamy-sand soils in Tifton, Georgia. In addition, numerous studies have reported that biochar application has only minimally impacted soil water retention [16][27][29], as fresh (i.e., young and unweathered) biochar is known to show some hydrophobicity, despite having large porosity, that could have prevented or restricted water from entering the soil pores, which has been observed in sandy soils [37][38]. Furthermore, aged (i.e., weathered) biochar has been reported to retain less water than fresh biochar in fine-loamy soils in Iowa [39].

Burke et al. [17] tracked estimated volumetric soil water contents at the 15-, 30-, and 45-cm depths over time at three locations in the eastern Arkansas portion of the LMRV. Burke et al. [17] reported soil water contents generally differed among biochar rates in the deeper soil depth, which was likely due to the study plot’s low BD and relatively low clay content (6%) that allowed furrow-irrigation water and rainwater to infiltrate to deeper depths where biochar had been applied. In the same plots as the soil samples were collected from for the current study, Burke et al. [17] reported that soil water contents varied among biochar-soil depth treatment combinations, where, at times, biochar application resulted in lower soil water contents than in the unamended control. In addition, biochar did not appear to maintain greater soil water contents at the 15-cm depth as the soil dried out from evaporation and a profile redistribution following furrow-irrigation events compared to the unamended control. Biochar effects appear to be minimized under irrigated conditions, thus are likely more beneficial for increasing soil water retention under dryland production systems [17].

There are several plausible explanations why there was no significant impact on water retention after three consecutive years of biochar application. Biochar has been reported to fill pores in sandy soils, which allows sandy soils to retain more water with biochar addition, due to biochar’s own porosity, than if biochar had not been applied [40]. However, biochar pores themselves can become clogged by silt and clay particles over time, preventing water from entering and being stored in biochar pores [41]. Biochar was minimally mixed into the upper few centimeters of soil after biochar application, which is not deep enough for biochar to have a substantial impact on soil water retention, with 10 cm being the suggested biochar mixing depth [42]. Furthermore, biochar application to reduced-tillage conditions is challenging, as biochar incorporation into fine-textured soils is greatly limited unless thoroughly mixed into the soil [43].

Though inconsistency exists, such as with the results of the current study and the minimal surface incorporation that occurred following annual biochar application, in general, biochar has been shown to increase soil water retention by various metrics, thus biochar is likely a useful organic amendment to impose positive effects for soil and water conservation. However, due to biochar’s potential to retain less water as the biochar ages, or decomposes and weathers in-situ [39][41], biochar may be better used as a short-term method of increasing soil water retention and be coupled with other conservation practices, such as reduced tillage and cover crops. In addition, similar to the results of the current study, biochar’s effectiveness at increasing soil water retention is likely soil-texture dependent, where positive biochar effects may manifest more in coarse- than fine-textured soils, as coarse-textured (i.e., sandy) soils already have lower water-retention capability, thus microporous biochar likely has a greater relative effect on soil water retention in sandy than in loamy or finer-textured soils [44].

Though even light incorporation may be challenging when a producer is also trying to maintain minimum or reduced tillage year after year, biochar incorporation into the 10 to 15 cm of soil profile is likely necessary to reduce potential off-site transport of the biochar via surface runoff during rainfall and/or irrigation events. Incorporation may also be necessary to maximize biochar’s ability to decrease the upper soil profile’s BD and increase soil porosity, infiltration, and water-retention capacity, particularly in regions, such as the LMRV, where many agricultural fields have a water-restricting, plow pan within 10 to 15 cm of the soil surface that limits vertical water flow after surface infiltration. Any added benefit that biochar could provide to improve vertical water flow with the soil profile after surface infiltration, such as through improving soil structure and aggregate-size distribution and increasing water-stable aggregation, will help off-set the decades long trend of declining groundwater recharge and increasing groundwater depletion due to a lengthy history of irrigated agriculture in the LMRV.