Comparison of Low-Cost Methods for Soil Water Holding Capacity

ABSTRACT The use of cost-effective methods for measurement of WHC is common in underdeveloped and developing countries, but the accuracy of these cost-effective methods compared to the sophisticated and more expensive alternatives is unclear. To compare different WHC measurement methods, 30 random samples of clay loam and sandy clay loam soils of Jhansi, India were used. The methods compared here were: FAO in-situ method (FAO), Keen-Raczkowski box method (KM), funnel method (FM), column method (CM) and pressure plate method (PPA). For WHC measurements the PPA results were comparable to KM and FM methods for sandy clay loam, and KM and FAO methods for clay loam. Therefore, until a reliable method that matches the results of sophisticated analytical methods of soil water measurement is available, different inexpensive analytical methods can be used, but they must be chosen with caution. The findings from this study will facilitate appropriate selection of a suitable method.


Introduction
Soil Water Holding Capacity (WHC), the amount of water a soil can physically retain against gravitational force, is one of the main indicators of soil quality and productivity. Soil WHC is also used as a key indicator of plant growth and microbial activity (Garcia, Hernandez, and Costa 1994;Krull, Skjemstad, and Baldock 2004). The soil WHC is influenced by many inherent factors, i.e., soil texture, structure, pore space, and organic matter content. It is well established that soils with finer particle size and higher organic matter hold more water than the soils with a coarse particle size and low organic matter. The soils with high WHC are less likely to leach nutrients and pesticides. For example, in a lysimeter study conducted in Germany, Meissner, Seeger, and Rupp (1998) reported that seep water volume was greater in sandy soil (150 mm) compared to clay soil (60 mm) and they attributed this to differences in soil WHC. Similarly, Mendes et al. (2015) found that potassium leaching was higher in sandy soil than in clay soil, due primarily to the low WHC of the sandy soil. Therefore, high soil WHC is vital for high agricultural production, irrespective of the climatic conditions.
Soil WHC is known to be influenced by certain agronomic management practices. For instance, changes in tillage regime may alter the bulk density or compaction of soil. Govindasamy et al. (2021) found that no-till applied on a clay loam soil in Texas, USA improved WHC over the conventional disking regime. Soil amendments such as biochar, manure, compost, or pumice have also been shown to alter WHC (Vengadaramana and Jashothan 2012;Verheijen et al. 2019). Malekian et al. (2012) reported that pumice addition to soil increased maize growth characteristics as a result of increased WHC. This indicates a positive relationship between soil WHC and crop growth and yield. Therefore, positive or negative impact of a management practice on WHC can directly indicate a potential positive or negative effect on yield, thus farmers/researchers should have access to a reliable and inexpensive method of measuring this important soil property.
There are numerous methods used to estimate soil WHC. The pressure plate method (PPA) is a standard technique. It's working matric potential range is −0.5 to −1.5 MPa (Cresswell, Green, and McKenzie 2008). However, the PPA method is expensive and time-consuming. Therefore, in developing countries, researchers favor inexpensive alternative methods, though they may be less accurate than the PPA method. Examples of commonly applied alternative methods include the funnel (FM) method (Bernard 1963;Govindasamy et al. 2021), the Keen Raczkowski box (KM) method (Keen and Raczkowski 1921), the column method (CM), and the FAO method (Motsara and Roy 2008). The FM, KM and CM are laboratory-based, whereas the FAO method is an in-situ procedure. Use of these methods to measure WHC is common, but the relationships and accuracy between WHC measured using different methods is not well understood for different soils (Motsara and Roy 2008;Wang et al. 2020). This could limit comprehensive interpretations, comparison of results across studies, and ultimately, the choice of measurement methods for WHC among researchers in developing countries (Roper et al. 2019).
It is important to compare and contrast the different methods available for WHC, for informing a suitable alternative method for a given soil environment, in order to save cost and efforts. The objective of this study was to assess and compare the accuracy of the different methods available for WHC measurement in two major semi-arid soils.

Soil sample collection
Soil samples were obtained from three different forage production systems: grassland (25°31′ 1.826′′ N, 78°33′9.19′′ E), a perennial production system (Bajra Napier hybrid; 25°31′ 11.83′′ N, 78°33′48.68′′ E), and an annual production system (sorghum; 25°31′ 28.22′′ N, 78°33′48.12′′ E) at the research farm, ICAR -Indian Grassland and Fodder Research Institute, Jhansi, India. In each system, a total of 30 random samples were collected using a core sampler (4.5 cm in diameter) for a depth of 15 cm. Later, 10 random samples were grouped into a single composite sample; thus, a total of three composite samples were collected from each production system. The soil type of the grassland and the perennial systems were sandy clay loam (25.14% sand, 53.86% silt, and 21% clay), with a bulk density of 1.25-1.28 Mg m −3 , a pH of 6.4-7.2, an EC of 0.05-0.12 ds m −1 , and an organic carbon content ranging from 0.69% to 1.02%. Whereas the annual system had a clay loam soil (35.14% sand, 43.86% silt, and 21% clay), with 1.28 Mg m −3 bulk density, 7.40 pH, 0.11 ds m −1 EC, and 0.62% organic carbon. Select physical properties of each soil are presented in Table 1.

WHC measurement
The WHC of the soil was determined using five well-known methods, i.e., FAO in-situ, the Keen box, the funnel, the column, and the PPA method. The PPA method was considered as the standard methodology.

FAO in-situ method (FAO)
The FAO in-situ methodology (Figure 1a) was developed by Motsara and Roy (2008). In this method, WHC is obtained after 48 to 72 hours of soil saturation. For this purpose, three 1 m -by-1 m size plots were selected. After removing weeds and residues, bunds/berms were formed around the plot and filled with enough water to completely saturate the soil. The plot area was then covered with a polyethylene sheet to prevent evaporation. After 24 hours of saturation, soil samples were collected in three replicates from the center of the plot at a depth of 15 cm. The water content was determined by gravimetric method for each day till the values of consecutive days were almost equal. Soil WHC is calculated using equation 1.
WHC of succeeding days a2, a3, etc., where, X is the weight of the empty moisture box, Y is the weight of the box with moist soil, and Z is the weight of the box with oven-dried soil. Daily reading was plotted on a graph and the lowest reading was considered as per the direction of Motsara and Roy (2008).

Keen raczkowski box method (KM)
The Keen Raczkowski box method was developed by Keen and Raczkowski (1921). In this method, a filter paper was placed at the bottom of the Keen Raczkowski box ( Figure 1b). Then, the soil was filled in the box by tapping the box 20 times using a wooden stick and continuously adding the soil until the box was filled up to the rim level. Finally, the top of the box was leveled by removing any surplus soil with a spatula. The box was then placed into a tray containing water (1.5 L) and left overnight. After a 24-hour saturation, the box was removed from the tray. The initial weight of the whole setup was recorded after wiping off the excess water at the bottom of the box with a filter paper. The final weight of the box was recorded after 72 hours of drying in the oven at 105℃. Soil WHC is determined using equation 2.
where a is the weight of the empty box + filter paper, b is the weight of box + filter paper + moist soil, and c is the weight of the box + filter paper + oven-dried soil.

Funnel method (FM)
The modified funnel method provided by Bernard (1963) was used for the determination of soil WHC. In this method, the soil samples were taken in a funnel (100 mL volume) with a filter paper placed at the bottom (Figure 1c). The funnel was mounted on the top of a graduated cylinder (500 mL in volume). Exactly 100 ml of water was added to each sample and left to drain for 72 hours. The amount of water collected in the measuring cylinder was recorded to determine WHC, equation 3.
where a is the amount of water added to funnel + soil, b is the amount of water collected in the flask after 72 hrs.

Column method (CM)
For this method, a plastic column with a height of 25 cm and an internal diameter of 4 cm was used ( Figure 1d). The bottom of the column was covered with a perforated polyethylene sheet. Soil was filled up to 20 cm height of the column by tapping 20 times with the help of a wooden stick. To saturate the soil, 200 ml of water was added to each column. Then, the packed columns were hung upside down on a stand. Readings of drained water were recorded after 72 hours. WHC is determined using equation 4.
where a is the amount of water added to column + soil, b is the amount of water collected in the flask after 72 hours.

Pressure plate method (PPA)
A pressure plate apparatus (Soil moisture Equipment Corp., Goleta, CA) was used to estimate the soil water content at field capacity (−1/3 bar or 0.033 MPa). Twenty-five gram of soil samples (2 mm size) were prepared in duplicates. The soil samples were filled in the retaining rings on the pressure plate cell. All the cells were covered with waxed paper and allowed to stand for 24 hours with excess water on the plate. Before assembling the cells into the extractor, the excess water was removed using a syringe. The pressure was adjusted to 0.003 MPa and after stopping the water flow due to soil water pressure equilibrium with air pressure. The samples were transferred immediately to the moisture box and the initial weight of the sample were measured. The samples were dried at 105℃ in an oven till constant weight (72 hour). Soil water content was calculated at field capacity using equation 5 (Chakraborty, Bandyopathyay, and Pradhan 2015).
Gravimetric moisture content % ð Þ ¼ Mbws À Mbds Mbds À Mb � 100 (5) where Mb is the weight of the empty moisture box, Mbws is the weight of the moisture box and wet soil, and Mbds is the weight of the moisture box and oven dry soil.

Data analysis
The data were analyzed using the statistical analysis software (SAS V9.3, SAS Institute, Cary, NC).
Comparisons of WHC among soil types (i.e., clay loam and sandy clay loam) were conducted separately for each methodology (i.e., Column, FAO, Keen box, Funnel, and PPA) using the mixed procedure (PROC MIXED) by keeping replication as the random effect. Mean separation between treatments was performed using Tukey's honestly significant difference test. The significance of all the data were tested at α = 0.05. To determine the relationship and to produce PPA values for the corresponding low-cost measurements a linear regression model was fitted for soil WHC measurements using PPA and four low-cost methods (CM, FAO, KM, and FM) for total (combined analysis of clay loam and sandy clay loam soils), clay loam, and sandy clay loam soils using PROC REG.

Comparisons among the low-cost soil water measurement methods
The mean water holding capacity values measured in this study were between 23.5% and 35% ( Table 2). The greatest water contents were observed with the CM (41.3% and 22.7%), KM (35% and 26.5%), and FM (34% and 22%) methods in clay loam and sandy clay loam soils, respectively, whereas the lowest content was observed with the FAO method (23.9% and 14.7%, respectively). Furthermore, the mean water content and the ranges between maximum and minimum measured water content by KM (35%, 33.3% and 37.4%, respectively) and FM (34%, 31.3%, and 36.4%, respectively) were similar for clay loam. However, three methods (i.e., CM, KM, and FM) performed similarly for sandy clay loam soil (

Low-cost methods and comparison with the pressure plate apparatus
Average water content and ranges between maximum and minimum measured water content by PPA (24.2% and 17.1%) and FAO (25.9% and 16.7%) methods were similar for both clay loam and sandy clay loam soils, respectively (Table 2). However, the most commonly used low-cost methods such as KM and FM measured 10% more water content than the pressure plate method in clay loam soil  ( Figure 2). Likewise, both the methods estimated 23% to 27% more water content in sandy clay loam soil compared to the pressure plate method (Figure 3).

Accuracy and relationship between different methods
Mean values of all measurements for the low-cost methods generally estimated greater WHC for both the clay loam and sandy clay loam soils compared to the PPA method. Linear regression between PPA and four low-cost methods are shown in Table 3 and Figure 4. Although different methods of measuring WHC have variable results, understanding relationships among results from the methods will be valuable in selecting an appropriate method. The two measures of WHC, PPA and the low-cost methods (CM, FAO, FM, KM) were highly related (Figure 4) for the total population of soils (combined analysis of both the soils), but the strength of the relationship varied with soil type (Table 3 and Figure 4). Separating results for individual soil revealed a stronger linear relationship between PPA and KM, PPA, and FM, and PPA and FAO for the sandy clay loam soil compared to the clay loam soil. The relationship for the clay loam soil was better for PPA and KM, and PPA and FAO which implies that PPA and KM measurement have similar relationships for clay loam soil. Among the low-cost methods, PPA and KM Table 3. Regression for soil water holding capacity measurements using pressure plate (PPA) and four low-cost methods (column (CM), FAO in-situ (FAO), keen box (KM), funnel (FM)) for total (combined analysis of both the soil type), clay loam, and sandy clay loam soils in the study.  (r 2 = 0.61) and PPA and FM (r 2 = 0.49), and FAO and PPA (r 2 = 0.38) were strongly related for sandy clay loam soil, but for clay loam soil PPA and KM (r 2 = 0.47), PPA and FAO (r 2 = 0.42) and PPA and FM (r 2 = 0.24) were the best. The produced regression equation (between low-cost methods and the PPA) also generates PPA values for a corresponding low-cost measurement value (Table 3). This is a major practical benefit for the scientist, who often uses low-cost methods such as KM, FAO, and FM methods.

Discussion
The mean water holding capacity values measured in this study were consistent with the water content for clay loam and sandy clay loam soils of previous studies (Basso et al. 2013;Viji and Prasanna 2012), which was typically between 23.5% and 35%. Among the low-cost methods, the laboratory methods (CM, KM, and FM) estimated a higher water content in clay and sandy clay loam soils compared to FAO in-situ method. This is presumably related to destruction of soil Figure 4 Linear regression model fitted for soil water holding capacity measurements using the pressure plate (PPA) and four low-cost methods (column (CM), FAO in-situ (FAO), keen box (KM), funnel (FM)) for total (combined analysis of both type of soil), clay loam and sandy clay loam soils in the study structure, smearing of soil macro pores and reduced influence of gravity under the laboratory conditions (Lekshmi, Singh, and Baghini 2014). The performance of laboratory methods varied with soil type, this may be due to differences in water absorption between clay loam and sandy clay loam soils. Clay loams generally absorb more water than the sandy clay loam soils (Little, Metelerkamp, and Smith 1998). Further, clay soils hold more hygroscopic water than sandy clay loams because of higher surface area (Lekshmi, Singh, and Baghini 2014). Moreover, clay soils dominated by micro-porosity may behave differently than macroporosity due to air entry in the field capacity region of the water release curve. Comparable results were observed between PPA and FAO methods, whereas most commonly used laboratory methods (KM and FM) were overestimated in both clay and sandy clay loam soils. One reason may be the use of pressure in PPA (applied pressure) and FAO (natural pressure), resulting in lower water content estimations in comparison with the other methods. Another possible reason may be the evaporative loss of water from the PPA and FAO methods due to applied pressure and solar radiation, respectively (Bechtold et al. 2018). Additionally, measurement errors are a function of porosity, specific surface, depth of the soil, pore fluid characteristics, degree of compaction, temperature, and humidity (Letey 1985;Wagner and Scheuermann 2009).
Generally, the low-cost methods estimated greater WHC for both the clay loam and sandy clay loam soils compared to the PPA method. Low-cost methods usually do not employ any additional pressure while estimating the soil water content which could explain the over estimation of water content. All the studied methods were comparable for total population of soil but the strength of relationship varied with soil type. A cost-effective conventional methods of measuring WHC have been used as an alternative to sophisticated instrument-based methods (Bhadha et al. 2017;Ghosh et al. 2021;Govindasamy et al. 2021). When using techniques other than the PPA and FAO measurements, the WHC content of soils has been reported to be as much as 70-78% (Bordoloi et al. 2019;Govindasamy et al. 2021), and as low as 30-45% in different soils (Bhadha et al. 2017;Bordoloi et al. 2019). The values reported here range from 9.4% to 43.2%. Although the relationships reported here are relatively weak (r 2 < 0.600), this is the first study to provide valuable information in quantifying the relationships between a reference method such as PPA and other low-cost methods in common usage. Furthermore, the produced regression equation (between low-cost methods and the PPA) also generates PPA values for a corresponding low-cost measurement value. This is a major practical benefit for the scientist, who often uses low-cost methods such as KM, FAO and FM methods.

Conclusion
This study is the first effort to compare several low-cost methods with pressure plate apparatus for WHC measurements. Variations in soil type result in discrepancies in soil water measurement when different methods are used. This can be explained by the physics of the specific measurement approach and the interaction of these approaches with the distribution of micro-and macro-porosity in soils of different textures. For instance, relationship between methods were more similar for sandy clay loam soil, but weaker for clay loam. Of the low-cost methods, keen box, funnel, and FAO methods were highly related to pressure plate method for sandy clay loam soil and for clay loam soil keen box and FAO methods were the suitable. Therefore, these three methods are likely to be the effective alternative method for evaluating WHC. As soil water content is more critical for crop management recommendations, it is important to note the limitations of existing low-cost methods and their sensitivity to detecting water levels under both field and laboratory conditions. Therefore, until a reliable method that matches the results of sophisticated analytical methods of soil water measurement is available, different inexpensive analytical methods can be used, but they must be chosen with caution.