Impact of Water Holding Capacity and Moisture Content of Soil Substrates on the Moisture Content of Wood in Terrestrial Microcosms

Abstract

Terrestrial microcosms (TMCs) are frequently used for testing the durability of wood and wood-based materials, as well as the protective effectiveness of wood preservatives. In contrary to experiments in soil ecology sciences, the experimental setup is usually rather simple. However, for service life prediction of wood exposed in ground, it is of imminent interest to better understand the different parameters defining the boundary conditions in TMCs. This study focused, therefore, on soil–wood–moisture interactions. Terrestrial microcosms were prepared from the same compost substrate with varying water holding capacities (WHCs) and soil moisture contents (MC_soil). Wood specimens were exposed to 48 TMCs with varying WHCs and MC_soil. The wood moisture content (MC_wood) was studied as well as its distribution within the specimens. For this purpose, the compost substrate was mixed with sand and peat and its WHC was determined using two methods in comparison, i.e., the “droplet counting method” and the “cylinder sand bath method” in which the latter turned out advantageous over the other. The MC_wood increased generally with rising MC_soil, but WHC was often negatively correlated with MC_wood. The distance to water saturation S_soil from which MC_wood increased most intensively was found to be wood-species specific and might, therefore, require further consideration in soil-bed durability-testing and service life modelling of wood in soil contact.

1. Introduction

For determining the durability of wood or the protective effectiveness of wood preservatives against soft rot fungi and other soil-inhabiting micro-organisms, terrestrial microcosms (TMCs) can be used. For this purpose, natural top soil or a fertile loam-based horticultural soil should be used and various requirements need to be fulfilled with respect to the soil substrate.

It is well known that many parameters affect the decay activity of soils [1,2,3,4,5]. Therefore, it is recommended to consider more than one in-ground field test site for durability testing of wood and more than one soil substrate for laboratory studies using TMCs [2,6,7,8,9].

Consequently, for standardized test protocols several parameters are more or less strictly defined. For instance, according to the European standard CEN/TS 15083-2 [10], the following soil-related boundary conditions need to be assured:

• pH 6–8
• no added agrochemical
• water holding capacity (WHC): 25–60%
• natural soils: peat or top 50 mm removed and not taken from a depth below 200 mm
• soil collected in moist conditions
• soil passed through a sieve of nominal aperture size 12.5 mm
• storage of soil prior to use only in closed moisture proof containers
• thorough mixing of soil before use
• horticultural soil which was sterilized during its preparation needs to be replenished with 20% natural soil
• soil not used before

Finally, a moisture content of the soil (MC_soil) equivalent to 95% of its WHC is required and the TMC should be stored at 27 °C ± 2 °C and 70% ± 5% relative humidity (RH) during the whole period of exposure in a dark room.

A previous study by Wälchli [11] showed that MC_wood decreased with both decreasing MC_soil and WHC as determined for two different soils and five different MC_soil. However, mass loss (ML) by decay of untreated and differently copper–chromium–boron (CCB)-treated Scots pine sapwood was neither correlated with MC_wood nor with MC_soil. Similarly, Mieß [12] found an increase in MC_wood with rising MC_soil in three different soil types and for different untreated and modified timbers. Furthermore, she found a gradient of MC_wood in untreated wood from the highest MC_wood in the bottom part and lowest MC_soil in the top part of the buried test stakes. In contrast, a remarkable 20% of the Scots pine sapwood specimens showed the highest MC_wood in the top or central part of the specimens. Mieß [12] suggested that the MC_wood gradients were the consequence of vertical gradients of MC_soil, which were differently severe due to the different soil wetting and re-drying regimes. It is further likely that the gradients were the consequence of ML gradients along the stake-shaped specimens, because the MC_wood of the different specimen segments had been determined not before the end of the test after 17 weeks of incubation when significant ML had already occurred.

Gray [13] performed durability tests in TMCs using different soils at different MC_soil and found that the highest ML occurred at an MC_soil between 108% and 148% of the WHC of the respective soil. The highest MC_wood after harvesting was found at an MC_soil between 120% and 218% of its WHC referring to an MC_soil at approximately 40% in all soil types used. Thus, ML increased with increasing MC_soil, but found an optimum, which was, however, far beyond the recommended 95% WHC. Again, MC_wood data are needed to obtain a set perspective, since they refer to the different severely decayed specimens after harvesting.

In summary, it becomes evident that both WHC and MC_soil influence MC_wood and ML through fungal decay, and do seemingly interact. Clear relationships between the three moisture-related parameters have not yet been established.

Others [6,12,14,15] previously demonstrated that all three rot types, i.e., brown, white, and soft rot, occur in TMCs complemented by tunneling, erosion, and cavity bacteria. However, neither MC_wood nor MC_soil seemed to limit their occurrence. Solely, soft rot apparently copes better with very high moisture contents, which are not favorable for brown and white rot fungi. Nevertheless, soft rot fungi can degrade wood in a rather large moisture range. They are early colonizers, so-called “ruderal organisms” [16], which, in contrast to basidiomycetes (‘combative organisms’), are rarely able to take over a substrate [17].

The WHC of soil substrates can vary remarkably, and therefore, it needs to be determined before each test. In both standards, CEN/TS 15083–2 [10] as well as ENV 807 [18], a suitable method for determining the WHC of soil is described: the so-called “droplet counting method”. The method is based on determining the ability of a sample of a test substrate to retain water against the pull of a vacuum pump, as a measure of its WHC. However, the method is rather laborious and time consuming. Furthermore, the standards lack a definition of the vacuum that needs to be applied, wherefore one might question the reproducibility of the test results.

Within this study, we conducted comparative WHC measurements on a series of different mixtures of compost and silica sand using the “droplet counting method” and an alternative method according to ISO 11268-2 [19], where wet soil samples are allowed to drain on a sand bath. Based on this comparison of methods, TMCs should be prepared representing soil substrates of varying WHCs and MC_soil. The overall objective of this study was to establish relationships between WHC, MC_soil, and the resulting MC_wood of different wood species after exposure in the TMC.

2. Materials and Methods

2.1. Soil Substrates

Three soil substrates were used to prepare TMCs of defined water holding capacities (WHCs). The basis substrate was a compost produced by the University of Goettingen from horticultural waste (i.e., leaf litter, grass, cut softwoods, and hardwoods, sand). To lower its WHC, silica sand (grain size > 0.2 mm) was added; to increase its WHC, peat (moderately-to-severely decomposed high-moor peat (H3–H8), total nitrogen 0.35%, magnesium 0.15%, organic substance 30%) was added. Both peat and compost were passed through a sieve of a nominal aperture size 8.5 mm. The soil moisture content (MC_soil) and the WHC were determined according to the “droplet counting method” and the “cylinder sand bath method”.

2.2. Determination of the soil Moisture Content (MC_soil)

Soil samples of 7–64 g (depending on the soil density) were taken for determining the soil moisture content (MC_soil). Three replicate samples were taken, weighed to the nearest 0.01 g, oven-dried at 103 °C, and weighed again. MC_soil was calculated as follows:

$${\mathrm{MC}}_{\mathrm{soil}}=\frac{{\mathrm{m}}_{\mathrm{soil},\mathrm{wet}}-{\mathrm{m}}_{\mathrm{soil},0}}{{\mathrm{m}}_{\mathrm{soil},0}}\times 100[\%]$$

(1)

where MC_soil is the soil moisture content, in %; m_soil,wet is the wet soil mass, in g; m_soil,0 is the oven-dry soil mass, in g.

2.3. Determination of the Water Holding Capacity (WHC) of Soil

2.3.1. “Droplet Counting Method”

A small quantity of water was added to soil samples of 200 g, the substrate was mixed well, and the operation was repeated until the soil particles stuck to another (crumb structure). Further, 25 mL water were added, mixed well, and allowed to stand for 2 h. A coarse filter paper was placed in the bottom of a Buchner funnel (100 mm diameter) and moistened to seal the filter paper to the funnel. The prepared test sample was transferred into the funnel and spread evenly. The bottom of the Buchner funnel was covered by soil substrate to a height of at least 10 mm. Suction was applied using a vacuum pump until no more than five drops of water per minute were being withdrawn from the sample, increasing the suction slowly to avoid perforation of the filter paper. The sample was transferred to an aluminum container of known mass and weighed. The container was oven-dried at 103 °C ± 2 °C and weighed again. The WHC of n = 5 peat samples and n = 3 sand and compost samples was determined and calculated as follows:

$$WHC=\frac{{\mathrm{m}}_{\mathrm{soil},\mathrm{saturated}}-{\mathrm{m}}_{\mathrm{soil},0}}{{\mathrm{m}}_{\mathrm{soil},0}}\times 100[\%]$$

(2)

where WHC is the water holding capacity, in %; m_{soil,saturated} is the soil mass at saturation, in g; m_soil,0 is the oven-dry soil mass, in g.

2.3.2. “Cylinder Sand Bath Method”

Soil was inserted into polyethylene cylinders with 4 cm diameters. The bottoms of the cylinders were covered with a fine polymer grid and filter paper (MN 640 W 70 mm). All cylinders were placed in a vat for 3 h, which was filled with water to a height 1 cm above the soil filling height of 7 cm. After soaking the soil in water, the cylinders were placed on a water-saturated sand bath for 2 h to allow the unbound water to drain. The soil samples were then weighed wet, oven-dried at 103 °C ± 2 °C, and the WHC of the soil was calculated according to Equation (2) analogously to the “droplet counting method”.

2.4. Preparation of Mixed Soil Substrates

For a comparison of the “Droplet counting method” and the “cylinder sand bath method” and for establishing a regression between mixing ratios of the different soil substrates and their resulting WHC, a total of 22 soil substrate mixtures were prepared as summarized in

Table 1. Mixing ratios of soil substrates for water holding capacity (WHC) tests. Percentage is based on the oven-dried mass.

Percentage Compost (%)	100	95	90	80	70	60	50	40	30	20	10
Percentage silica sand/peat (%)	0	5	10	20	30	40	50	60	70	80	90

.

For preparing mixed soil substrates, the following equation was used:

$$m_{soil x,target, wet}=m_{target, total, 0}\times (\frac{a_{target,0}}{100})\times (\frac{100}{100-M{C}_{soil x}})(\mathrm{g})$$

(3)

where m_{soil x,target, wet} is the target mass of the wet substrate x, in g; m_{target, total, 0} is the target oven-dry mass of the total mix, in g; a_target,0 is the target percentage of substrate x based on the oven-dry mass, in %. MC_{soil x,} is the soil moisture content of substrate x, in %.

2.5. Terrestrial Microcosms (TMCs)

Miniature terrestrial microcosms were prepared in polypropylene containers of 110 (height) × 110 × 80 mm³ and a volume of 500 mL. In total, 48 different substrates were filled in the containers each to a height of 100 mm. The combinations of the parameters WHC and MC_soil are summarized in

Table 2. Soil moisture content MC_soil (%) for combinations of target WHC ¹ and target MC_soil expressed as (%WHC).

MCsoil (%WHC) 1	WHC (%) 1
	21.8 2	30.0	40.0	50.0	60.0	70.0	80.0	90.0
30.0	6.5	9.0	12.0	15.0	18.0	21.0	24.0	27.0
50.0	10.9	15.0	20.0	25.0	30.0	35.0	40.0	45.0
70.0	15.3	21.0	28.0	35.0	42.0	49.0	56.0	63.0
80.0	17.4	24.0	32.0	40.0	48.0	56.0	64.0	72.0
95.0	20.7	28.5	38.0	47.5	57.0	66.5	76.0	85.5
120.0	26.2	36.0	48.0	60.0	72.0	84.0	96.0	108.0

¹ WHC determined according to the “cylinder sand bath method” according to ISO 11268-2 [19]. ² WHC of pure silica sand was 21.8% and consequently the lowest WHC achieved.

, where the latter is expressed as (%WHC). The containers were weighed to the nearest 0.01 g, closed with a lid, and their total mass maintained over a period of three weeks.

The WHC of the substrates used for TMCs were determined exclusively according to ISO 11268-2 [19]. The basic substrate was compost. Silica sand and peat were added according to the regression obtained by comparative WHC measurements as described in Section 2.4 (Figure 1).

The following regression functions were used (see also Section 3.1):

$$WH{C}_{mix:compost-sand}=-0.586\times a_{target, sand, 0}+80.81(\%)$$

(4)

$$WH{C}_{mix:compost-turf}=2.499\times a_{target, turf, 0}+87.15(\%)$$

(5)

where WHC_{mix:compost-sand} is the water holding capacity of compost mixed with silica sand, in %; WHC_{mix:compost-peat} is the water holding capacity of compost mixed with peat, in %; a_{target, sand, 0} is the target percentage of sand based on the oven-dry mass, in %; a_{target, peat, 0} is the target percentage of peat based on the oven-dry mass, in %.

The soil mixtures used for the TMCs are summarized in

Table 3. WHC of different mixtures of compost with sand and peat.

	WHC (%) 1
	21.8 2	30.0	40.0	50.0	60.0	70.0	80.0	90.0
Percentage compost (%)	0.0	13.2	30.3	47.4	64.5	81.6	98.6	98.6
Percentage sand (%)	100.0	86.8	69.7	52.6	35.5	18.5	1.4	0.0
Percentage peat (%)	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.4

¹ WHC determined according to the “cylinder sand bath method” according to ISO 11268-2 [19]. ² WHC of pure silica sand was 21.8% and consequently the lowest WHC achieved.

.

2.6. Preparation and Exposure of Wood Specimens

Specimens of 5 × 10 × 100 (ax.) mm³ were prepared from Scots pine sapwood (Pinus sylvestris L.), Douglas fir heartwood (Pseudotsuga menziesii Franco), English oak heartwood (Quercus robur L.), and European beech (Fagus sylvatica L.). All specimens were free from defects such as cracks, decay, and discoloration. For each of the 48 combinations of WHC and MC_soil, n = 5 replicate specimens of each species were prepared, which corresponded to a total of 960 specimens.

In total, 48 soil substrates, i.e., combinations of WHC and MC_soil, were prepared and each was used to fill two containers (miniature TMCs). Wood specimens were conditioned at 20 °C/65% RH until constant mass before soil exposure. Afterwards, ten wood specimens were buried to 4/5 of their length in each container and exposed for three weeks. The MC_soil was maintained by adding water about every third day if needed.

2.7. Determination of the Wood Moisture Content (MC_wood)

Specimens from selected TMCs were used to determine the MC_wood distribution within the specimens. Therefore, after harvest, the specimens were cleaned from adhering soil particles, cut into five segments of 20 mm length using ratchet scissors, weighed, dried, and weighed again in each step separately. Segment-wise MC_wood was determined on specimens after exposure in TMC with substrates of 30, 60, and 90% WHC, and a MC_soil of 30, 75, and 95% of its WHC.

$${\mathrm{MC}}_{\mathrm{wood}}=\frac{{\mathrm{m}}_{\mathrm{wood},\mathrm{wet}}-{\mathrm{m}}_{\mathrm{wood},0}}{{\mathrm{m}}_{\mathrm{wood},0}}\times 100(\%)$$

(6)

where MC_wood is the wood moisture content, in %; m_wood,wet is the wet mass of the wood specimen, in g; m_wood,0 is the oven-dry wood mass, in g.

2.8. Statistical Analysis

Regression functions between different variables were established using the method of least squares to achieve the best fit. Statistical differences between collectives were considered significant at a probability of error less than 5% according to a modified Student t-test (Welch test).

3. Results and Discussion

3.1. Water Holding Capacity (WHC) of Soil Mixtures

The WHC of the three initial substrates was highest for peat, followed by compost and sand according to both methods applied (

Table 4. WHC (%) of the initial soil substrates determined according to the “droplet counting method” and the “cylinder sand bath method”. Standard deviation in parentheses.

	WHC (%) 1
	Silica Sand	Compost	Peat
Droplet counting method	6.7 (0.3)	78.9 (1.3)	296.6 (40.9)
Cylinder sand bath method	21.8 (0.4)	75.5 (1.1)	318.3 (38.8)

¹ Number of replicate samples was n = 3 and n = 5 for peat tested according to the droplet counting method.

).

According to both methods, the WHC of the different soil mixtures was linearly correlated with the percentage of added sand or peat, respectively. With increasing percentage of sand and decreasing percentage of peat, the WHC decreased (Figure 1).

It became evident that: (1) single WHC values scattered more and (2) the regression between substrate ratios and WHC was less pronounced when using the “droplet counting method”. Furthermore, the following advantages of the “cylinder sand bath method” over the “droplet counting method” became apparent:

• Less time consumption: 7 min/sample compared to more than 60 min/sample using the “droplet counting method”.
• Consistency of the test setup: Time for counting droplets varied between 6 and 60 min/sample. Intervals between single droplets varied partly drastically, especially when testing substrates of high WHC. In contrast, up to several hundred cylinders can be used in parallel and the duration of test was always constant.
• Clear description of setup: Simple, less expensive, and well described setup. In contrast, the description of the “droplet counting method” suffers from some vagueness: size of Buchner funnel, applied vacuum, and type of filter paper are not specified, but likely affect the test results.
• Independency from sample size and volume: In contrast, the given sample thickness according to the “droplet counting method” led to strongly varying mass of the soil sample in the Buchner funnel and affected the WHC, as shown for peat in Figure 2. With increasing sample size (= sample mass) the WHC increased.

Generally, it was observed that substrates with very high WHC, such as peat, require longer wetting periods than specified by the standards, i.e., 1–2 h according to CEN/TS 15083-2 [10] and 3 h according to ISO 11268-2 [19]. After 3 h of submersion, the peat was still not fully water saturated when oven-dried before, which consequently led to an underestimation of its WHC.

The WHC determined according to both methods were highly correlated, especially for WHC below 200%, as shown in Figure 3. Therefore, and regarding its numerous advantages, in the following, all WHC measurements were conducted using the “cylinder sand bath method”.

3.2. Impact of WHC and MC_soil on the Moisture Content of Wood (MC_wood) Exposed in TMCs

After three weeks of exposure in different TMCs, the average MC_wood was highest in Scots pine sapwood (88%), followed by English oak (75%), Beech (67%), and Douglas fir (48%). In general, MC_wood increased with increasing MC_soil, but was strongly dependent on the WHC of the soil. The lower the WHC, the higher was the MC_wood at a given MC_soil, which coincided with previous findings [12]. Lower WHC in this study corresponded with a higher percentages of silica sand, which can only physically absorb water in contrast to organic soil substrates such as compost soil and peat, which also form chemical bonds with water [20]. The capacity to bind water is therefore higher in organic substrates which restricts the amount of available water which potentially wets the wood specimens.

This effect became especially prominent when considering MC_soil as a percentage of the WHC of the soil as illustrated in Figure 4. The lower MC_soil (%WHC), the lower the MC_wood was at a given WHC of the substrate in the TMCs. However, it also became apparent that the effect of MC_soil became more pronounced at higher WHCs, i.e., the range of MC_wood between 30% and 120% MC_soil expressed as percentage of its WHC was higher by up 2 factors in substrates of high WHC (120%) compared to those with very low WHC (30%).

The difference between MC_wood results achieved after three weeks of in-soil exposure was surprisingly small between Beech and English oak heartwood, because the latter is known to take up water more slowly due to the formation of tyloses in the vessels. In contrast, Beech wood—apart from false heartwood which was excluded in this study— usually takes up liquid water very easy, although its vessel diameters are much smaller compared to the early wood vessels of English oak. Similarly, the maximum MC_wood of Douglas fir heartwood was in the same range of that of Scots pine sapwood when exposed in soil at an MC_soil of 120% WHC. Solely, at a lower MC_soil, the more permeable Scots pine sapwood showed higher MC_wood compared to the refractory heartwood of Douglas fir. In summary, it became evident that already after a short exposure period of three weeks in wet soil, wood anatomy-induced differences in moisture uptake diminished confirming previous findings [12].

To further illustrate the interdependency between WHC and MC_soil and their effect on MC_wood, the distance to water saturation of the soil substrate (S_soil) was determined according to Equation (7) and correlated with MC_wood (Figure 5).

$$S_{soil}={\mathrm{MC}}_{\mathrm{soil}}-\mathrm{WHC}(\%-\mathrm{points})$$

(7)

where S_soil is the distance to water saturation of the soil substrate, in %-points; MC_soil is the soil moisture content, in %; WHC is the water holding capacity, in %.

Generally, with increasing S_soil, the MC_wood increased as well, but followed wood species-specific curves with differently steep increases and a plateau at S_soil = 0%, i.e., soil water saturation. For English oak, Beech, and Douglas fir, the MC_wood stayed approximately constant at S_soil > 0%. Solely, for Scots pine sapwood, MC_wood dropped significantly after exceeding the saturation point, although unlimited uptake of liquid water was expected to be provided above this threshold.

The distance to water saturation from which MC_wood remarkably rose differed also between wood species and was approximately at −55%-points for English oak, −25%-points for Beech and Douglas fir, and −20%-points for Scots pine sapwood. Scots pine sapwood showed also by far the highest increase in MC_wood with increasing S_soil, i.e., between 32% and up to 180% MC_wood between −20 and 0% S_soil.

3.3. Moisture Content Gradients in Buried Wood Specimens

The MC_wood in specimens buried to 4/5 of their length in TMCs showed partly drastic gradients from high moisture content in the bottom to less in the upper part, which was not buried (Figure 6, Figure 7, Figure 8 and Figure 9). Solely, Scots pine sapwood specimens exposed at high MC_soil (95%WHC) showed barely significant gradients, but very high MC_wood in all parts of the specimen. Similarly, deviating MC_wood gradients were reported by Mieß [12] for Scots pine sapwood specimens. As expected, generally, the highest difference in MC_wood was found between the upper segments and upper next segments.

The MC_wood in the upper segments of English oak and Douglas fir specimens was in the range of their equilibrium moisture content (EMC) at fiber saturation. In contrast, the upper segments of Scots pine sapwood and Beech specimens showed MC_wood up to 180 and 80% respectively, indicating strong capillary water transport along the specimen axis.

The MC_wood gradient in the specimens was positively correlated with MC_soil (%WHC). However, two types of MC gradients became apparent: (1) increasing MC_wood from the upper to the next segment, but rather constant MC_wood below (e.g., English oak at WHC = 30%), and (2) a steady increase of MC_wood from the upper to the bottom segment (e.g., Beech and Scots pine sapwood).

4. Conclusions

The findings from this laboratory study on the soil–wood–moisture interactions in terrestrial microcosms led us to the following conclusions:

• The more advantageous “Cylinder sand bath method” should consequently be seen as an adequate alternative for the “Droplet counting method”, which turned out disadvantageous regarding practical applicability, reproducibility, and reliability.
• Water holding capacityvalues obtained from both test methods applied seemed to be easily transferable to each other. It is, therefore, recommended to replace the “droplet counting method” with the “cylinder sand bath method”.
• The average MC_wood of specimens buried in TMCs increased with rising MC_soil, but WHC was often negatively correlated with MC_wood.
• The distance to water saturation S_soil appeared as a more predictive measure for MC_wood.
• With increasing S_soil the MC_wood increased but followed wood species-specific curves with differently steep increase and a plateau at S_soil = 0%.
• The distance to water saturation S_soil from which MC_wood increased most intensively was found to be wood-species specific and might, therefore, require further consideration in soil-bed durability testing. Thus, S_soil can likely be used to establish moisture conditions which are favorable for a specific decay type, i.e., brown, white or soft rot.
• The segment-wise determination of MC_wood revealed that a combination of low MC_soil and high WHC of the soil can easily lead to moisture conditions which are not favorable, neither for fungal decay in general, nor for soft rot decay in particular. They should, therefore, be avoided in durability testing, but might be of interest for the service life prediction of wood exposed to soil.

Based on the findings from this study, further experiments have been initiated to examine the effect of soil–wood–moisture interactions on fungal decay, in particular soft rot decay. In addition, the outstanding role of soil and wood temperature on decay in constantly wet wood will be further investigated.