ASSESSING THE NATURAL DURABILITY OF DIFFERENT TROPICAL TIMBERS IN SOIL-BED TESTS

Ground contact speeds up timber decay because of the large number of microorganisms in soil. This study, we assessed the natural durability of seven tropical species using the European standard EN-807 (2001). We embedded samples of Dalbergia granadillo, Cordia elaeagnoides, Swietenia humillis, Tabebuia donell-smithii, Hura polyandra, Enterolobium cyclocarpum and Tabebuia rosea and temperate species Fagus sylvatica (as a control) in sandy, clay-sandy-loam and clay-loam for 8, 16, 24 and 32 weeks. We evaluated durability of the samples by determining the mass loss and modulus of elasticity (MOE) loss. The results varied significantly (p < 0,001) depending on timber species and soil type considered. The D. granadillo and C. elaeagnoides were the most durable, with mass losses of 4,5%; 6,5% and MOE losses of 4,5%; 20,5%; respectively. F. sylvatica, T. rosea and E. cyclocarpum samples were the least durable, with mass losses of 22,3-25% and MOE losses of 35,8-59,8% respectively. Decay was most aggressive in sandy-clay-loam soil followed by the clay-loam soil and finally the sandy soil.


INTRODUCTION
Tropical timber species vary in color, chemical composition, physical and mechanical properties and durability.Some woods are particularly attractive and is used to make furniture, musical instruments, car interior trims and various types of handicrafts.Worldwide trade in tropical timber products reached a global volume of 273,1 million m 3 in 2014 (OIMT 2015).According to the Statistical Yearbook of Forestry Production of 2015 (SEMARNAT 2016), 336064m 3 (Total Stem Volume, TSV) of common tropical timber species and 13378 m 3 TSV of precious tropical species were harvested in Mexico, represent 5,7% of the total production and with a value of US$ 24909087,1 (SEMARNAT 2016).
Technological characterization of lesser known tropical species (LKTS) have enabled replacement of over-harvested timber species with some of these other species in the furniture-making and musical instrument sectors.In addition to aesthetic criteria used for timber species selection, natural durability is one of the most important attributes.
Natural durability is defined as the resistance of the timber species to degradation by biotic and abiotic media (Eaton and Halle 1993).Durability can be evaluated using a variety of laboratory and field-based methods (Meyer et al. 2014).The most common method involves embedding the timber in soil, which after sea water, is one of the most aggressive types of environment to which timber is exposed.The wide variety of microorganisms in soil, such as soft rot fungi (Ascomycetes) guarantee the effectiveness of the test, because microorganisms are capable of adapting to different environmental conditions (soil) and to timber preservatives (Schmidt 2006).The rate of decay of timber varies according to the type of soil used.Factors such as soil type, moisture content, chemical properties and the microorganisms present influence the type of attack and level of degradation (Jebrane et al. 2014, De-Avila et al. 2016).
The main objective of this study was to evaluate the natural durability of samples of the tropical timber species Cordia elaeagnoides, Dalbergia granadillo, Enterolobium cyclocarpum, Hura polyandra, Swietenia humillis, Tabebuia donell-smithii and Tabebuia rosea, were loss of mass and modulus of elasticity caused by exposure of the timber samples to different types of soil under laboratory conditions.

Timber species and sample preparation
Samples of undamaged, defect-free heartwood from Cordia elaeagnoides DC, Dalbergia granadillo Pittier, Enterolobium cyclocarpum (Jacq.)Griseb, Hura polyandra Baill, Swietenia humillis Zucc, Tabebuia donell-smithii Rose and Tabebuia rosea (Bertol) DC and of the temperate specie Fagus sylvatica L. (as a control) were obtained from the Faculty of Wood Technology, University of San Nicolás de Hidalgo (México).The wood was cut into 100 x 10 x 5 mm 3 (length x radial x tangential) specimens for testing.

Soil origin, preparation and analysis
Clay-sandy-loam, clay-loam and sandy soils were collected from sites in the Mexican states of Veracruz, Nuevo León and Michoacán.The samples were placed in plastic containers (60 x 40 x 15 cm 3 ) for the test and maintained at 20±2 °C; humidity, 65% ±2; and soil moisture content, 95% of water holding capacity (WHC).The soil pH, electrical conductivity and texture were determined according to NOM-021-RECNAT (2000), and the organic matter content following Walkley and Black (1934).The site description and analytical results are summarized in Table 1.
Table 1: Characteristics of the clay-sandy-loam, sandy and clay-loam soils collected from Veracruz, Michoacán and Nuevo León (Mexico) used to determine the natural durability of samples of seven tropical timber species.

Determination of natural durability of timber
Timber durability was determined on the basis of European standard EN-807 (2001).Twelve samples of each timber species were set aside as control (week zero).Another forty-eight samples of each species were partially buried in containers of each type of soil.The samples were embedded at random positions in the soil to 80% of their length.Each sample was separated by a space of 20 mm.Twelve samples of each timber species were extracted from each type of soil after 8, 16, 24 and 32 weeks.All samples were conditioned (20 °C and 65% relative humidity) to 12% moisture content, before the modulus of elasticity (MOE) test was applied.

Mass loss
The mass loss was determined gravimetrically at the end of MOE test with Equation 2. where: • Wl = mass loss (%) • W 1 = mass at the beginning of the test (g) • W 2 = mass at the end of the test (g)

Statistical analysis
The mass and MOE loss data were subjected to analysis of variance (ANOVA) with a factorial arrangement of 8 x 3 x 5.The factors were wood species (eight), soil type (three) and exposure time (five), with twelve repetitions for each treatment.Tukey's test was applied when the ANOVA indicated a statistically significant difference between the factors.All statistical analyses were performed with SPSS version 20.

Mass loss
The timber mass decreased from on average 22,9% to 4,1% after the samples had been in contact with the different types of soil for 32 weeks.The ANOVA revealed highly significant differences (p < 0,0001) for species, contact time and type of soil, and for the contact time x soil, species x soil, and species x contact time interactions (Table 2).Tukey tests for differences between species, soil type and contact time revealed statistically significant differences (Table 3, Table 4 and Table 5).

MOE loss
Average MOE values decreased between 9,4% and 56,7% after the samples were embedded in soil for 32 weeks.ANOVA of the MOE loss data revealed highly statistically significant differences (p < 0,0001) for species, contact time and type of soil, as well as for the species x contact time and species x soil type interactions (Table 2).Tukey tests for species, soil type and contact time revealed significant differences (Table 3, Table 4 and Table 5).
Table 6: Mean mass loss in samples of seven tropical timber species and the control species Fagus sylvatica embedded for 8, 16, 24 and 32 weeks in three different types of soil.
Table 7: Mean MOE loss in samples of seven tropical timber species and the control species Fagus sylvatica embedded for 8, 16, 24 and 32 weeks in three different types of soil.

Durability of timber
The most durable timbers were C. elaeagnoides and D. granadillo, with mean values of mass loss in the different types of soil ranging from 4,1% to 4,3% (Table 3 and Table 6).The natural durability of both timbers were similar to that of Tectona grandis reported as highly durable after application of a similar test (Machek et al. 2001).Tests of tropical and European timber species also revealed some highly durable types of timber, with mean mass loss of 3% in Acacia nigrescens, Pericopsis angolensis, Icuria dunensis and Pseudolachnostylis maprounaefolia, and some species with low durability, with mean mass loss between 40% and 48% (Ali et al. 2011).The latter values are similar to those obtained in the present study for the non-durable timbers species.High durability is attributed to the effect of extractives Hillis (1972), Tsoumis (1991), Taylor (2004), Syofuna et al. (2012), Kadir (2017), as well as to the high density that causes reduction of microbial growth due to lack of oxygen (Tsoumis 1991).The T. donell-smithii, S. humillis and H. polyandra samples were of moderate durability (Table 3), with mass loss values ranging from 6,2% to 9,9% (Table 6).The T. rosea, E. cy-clocarpum and F. sylvatica (control) timber samples were the least durable, with mass losses of 20,6%; 21,6% and 22,9% respectively (Table 6).The low durability of these timbers is attributed to low amounts of extracts (Taylor 2004).

Effect of soil type
The mass loss and MOE loss of the timber samples differed significantly depending on soil type (Table 4, Table 6 and Table 7).The sandy-clay-loam soil from Veracruz yielded the highest mass and MOE losses: mean values of 17,6% and 33,9% respectively.The sandy-loam soil from Nuevo León state yielded mean mass and MOE losses of 12,3% and 29,7% respectively.The sandy soil from Michoacán state produced mean mass and MOE losses of 9,8% and 22,5% respectively.The high mass and MOE losses caused by the loam-clay-sandy soil are due to the high pH (7,8) and high organic matter content (3,4) (Table 1).
In tests carried out with timber of low durability (F.sylvatica, Populus spp.and Ulmus spp.) mass losses values ranged between 35% and 40%, and MOE losses as well between 77% and 86% (Machek et al. 2001); similar values were obtained in the present test for F. sylvatica, E. cyclocarpum and T. rosea (Table 6 and Table  7).
Differences in mass loss in relation to soil type were also observed by Nami et al. (2006), Brischke et al. (2009) and by De-Avila et al. (2016) who concluded that mass losses, even from impregnated timber, are related to the type of soil used for the test.The sandy-clay-loam soil from Veracruz with pH 7,8 and an organic matter content of 3,4 % caused the highest level of deterioration in the samples of tropical timber tested.The maximum degradation of timber reported by Gersonde and Kerner (1984) was yielded by soil of pH 5,5-5,8 and organic matter content of between 5,8%-10%.
The speed of timber species degradation in terrestrial microcosm systems is related to the soil physical and chemical characteristics (Bravery 1975, Gersonde and Kerner 1984, Leightley and Russell 1980).However, these authors did not observe any correlation between the organic matter content and pH.They also reported that as a medium for timber degradation experiments, soil represents a complex system in which it is difficult to consider all variables involved, as the characteristics of different types of soil act and react in many different ways during the rotting process.The fungi and bacteria in the soil substrate is the dominating factor that influences the decay potential of the soil (Brischke et al. 2013) as well as the higher water holding capacity (Jebrane et al. 2014) while low oxygen availability is likely to reduce decay hazard (Wakeling 2006).Thus, while Savory (1955) noted a direct relationship between the soil organic matter content and timber rotting, in which mass loss is directly proportional to the organic matter content, the nitrogen content and aeration also affected the level of degradation of soil.

Effect of contact time
The timber durability, as indicated by both mass loss and MOE loss, varied in relation to exposure time (Table 5).The mean values of mass loss and MOE loss after the samples had been embedded in the soil for 32 weeks were 17,1% and 48,3% respectively.The MOE loss was the best indicator of the durability of timber species tested, as it was observed at early stages of decay, even when mass loss was low (Liese 1955, Hardie 1980, Ross and Pellerin 1994, Machek et al. 1997).Intensive degradation of carbohydrates by soft rot fungi produces MOE loss of 50% and only 5% of mass loss, thus causing the wood to crack as a result of reduced dimensional stability (Schmidt 2006).This was also observed in the samples of tropical timbers tested in the present study.

Table 2 :
Summary of the ANOVA results for mass loss and modulus of elasticity (MOE) loss in samples of seven tropical timber species embedded for 32 weeks in three different types of soil.

Table 3 :
Mean mass and modulus of elasticity (MOE) loss in seven tropical timber embedded for 32 weeks in three soil types.Significant difference in the mean values in relation to tree species are indicated by different lower case letters (Tukey test P<0,05).

Table 4 :
Mean values of mass and modulus of elasticity (MOE) loss in samples of seven tropical timber species embedded for 32 weeks in three different types of soil.Significant differences in relation to soil type are indicated by different lower case letters.

Table 5 :
Mean mass and modulus of elasticity (MOE) losses in samples of tropical timbers exposed for different times in three different soils.Significant differences in relation to time of exposure are indicated by different lower case letters (Tukey test, p<0,05).