Introduction

The restoration of forest ecosystems after mining operations is the goal of community groups, government agencies, and mining companies worldwide. In Australia, an important aspect of this restoration is that mined ecosystems are integrated back into routine forest management regimes, including stand thinning and fuel reducing prescribed fire [23]. Therefore, considerable research effort has been directed toward the reestablishment and management of vegetation communities in mined forests [25–27]. However, few studies have examined the functioning of soil microbial communities after bauxite mining [41, 62], although the microbial community fulfills critical ecological roles such as decomposition and nutrient cycling.

In Western Australia, forests established on sites previously mined for bauxite often have high stand densities owing to heavy seeding or planting regimes that are designed to achieve rapid vegetation cover [20]. However, tree density on rehabilitated sites is much greater than that of the surrounding forest [24, 55] and can lead to understorey suppression, decreased growth, and increased water use. Consequently, manual thinning is necessary to increase productivity of individual trees, reduce total leaf area, and increase water yield within catchment areas [23]. Thinning operations transfer nutrients from aboveground biomass to the soil surface [21] and opens the forest canopy so that greater quantities of water and sunlight reach the forest floor. Green litter from thinned trees left on site is likely to contain higher concentrations of nutrients that are more readily decomposed than litter returned to the forest floor after senescence [21, 50]. Conversely, slash (the woody residue left on ground after harvesting) is likely to be more recalcitrant than litter dominated by foliage, as it contains less nutrients and has much greater carbon/nitrogen ratios [50]. Therefore, the quantity and quality of organic substrates presented to the soil microbial community within thinned and nonthinned forests may vary widely. As the individual species, which comprise the microbial community, have different capacities to respire different organic substrates [6, 16], forest thinning may give rise to a soil microbial community that is distinct from that of nonthinned soil. In turn, the subsequent ability of these microbial communities to respire a wide variety of added organic substrates may also differ.

High stand density and annual biomass production in rehabilitated Jarrah forest also produce large fire fuel loads [44]. Therefore, after 12–15 years of rehabilitation, sites are prescribed burned. The impacts of fire on soil properties in mined forest ecosystems will vary extensively but will depend largely on the intensity of the fire [39, 46]. High intensity fires can result in the net loss of soil nutrients, particularly of nitrogen (N) via volatilization, ash transport, and runoff [39, 52]. In contrast, low intensity fires may increase nutrient availability via ash deposition and soil heating [8, 32, 69]. In rehabilitated Jarrah forest, the intensity of prescribed fires during autumn (270–20,300 kW m−1) have been reported to be considerably greater than prescribed fires during spring (188–2,080 kW m−1) [25]. Consequently, restoration ecologists have recommended burning under low intensity spring conditions [46]. However, low intensity burns still combust biomass and surface organic matter, which are likely to produce short term (<1 year) changes in the quality and quantity of organic matter presented to soil microbes. Therefore, prescribed fire may also give rise to distinct microbial communities that differ in their ability to decompose a variety of added organic substrates compared with nonburned soils.

Mined sites are often contour ripped before seeding to alleviate soil compaction, assist tree root penetration, maximize water infiltration, and assist in controlling erosion [63]. The distinct undulations (≈0.5 m wide and 0.25 m deep) that result from ripping have significant implications for subsequent plant litter distribution and are a major source of spatial heterogeneity in soil nutrients [63]. Soil from within furrows is often characterized by a larger organic matter gradient between surface (0–5 cm) and subsurface (5–10 cm) soil, relative to soil from mounds [67]. Subsequent forest thinning is likely to increase organic matter heterogeneity as greater quantities of organic matter fall to the soil surface in the short-term relative to nonthinned sites. Alternatively, low intensity fire will combust litter and possibly decrease organic matter heterogeneity relative to nonburned sites. Furthermore, any alteration of soil chemical, physical, and biological properties caused by soil heating during burning are also likely to be greatest in surface than subsurface soil [7]. As the availability of organic substrates is likely to mediate microbial function, soil depth, forest thinning, and/or prescribed fire are also more likely to affect microbial substrate utilization in the substrate-rich furrows than substrate-poor mounds [41].

The specific aims of this study were to (1) assess the degree to which forest thinning, prescribed fire, contour ripping, and soil depth influence community level physiological profiles (CLPP) of the microbial community and (2) examine the relationships between microbial CLPP and selected soil properties. We proposed that (1) forest thinning, prescribed fire, contour ripping, and soil depth would produce microbial communities with distinct substrate utilization patterns; (2) differences in substrate utilization between mound/furrow and soil depths would be greater in thinned and nonburned soil than nonthinned and burned soil; and (3) microbial substrate utilization is mediated by the quality and quantity of organic substrates present in the soil.

Materials and Methods

Soil Description and Treatments

The study site was located approximately 110 km south of Perth, Western Australia, in a Jarrah (Eucalyptus marginata Donn ex Sm.) forest that had been mined for alumina contained in bauxite and undergone initial rehabilitation in 1992. A detailed description of the procedures used to rehabilitate bauxite mines in south-west Western Australia are described elsewhere [20, 55, 67]. Standard procedures include ripping along contours (≈0.5 m wide and 0.25 m deep) to produce distinct mounds and furrows. At our study sites, three treatment blocks (3 ha) were thinned in April (autumn) of 2003 from 3,000–8,000 stems ha−1 to a density of 600–800 stems ha−1 and/or aerially burned in September (spring) 2003. Sites that were not thinned or burned were also established adjacent to thinned and burned sites. The dominant overstorey plant species were Jarrah and Marri (Corymbia calophylla; seeded at a ratio of 4:1) with an understorey consisting of a variety of native legumes and nonlegumes.

Three replicate plots (20 × 20 m) were selected within thinned, burned, thinned/burned, and nonthinned/nonburned sites to encompass site variation in understorey composition, productivity, and burning intensity. In August 2004, we collected separate composite soil samples from 12 positions within either mounds or furrows across each 20 × 20-m plot. Samples were bulked by depth (0–5 cm and 5–10 cm). All soils were sieved (<2 mm) and stored at 4°C for less than 2 weeks before analysis.

Coarse and Fine Litter

Litter was collected as coarse and fine material from each of three 4 × 4-m subplots established in alignment with contour ripping lines within each treatment plot. The size and alignment of these plots was specifically chosen to encompass natural spatial variation in soil characteristics that result from rip lines [63]. Mound litter was collected from a 0.50 × 1-m quadrat centered and running along the middle of a mound in each subplot. Furrow litter was collect from a 0.50 × 1-m quadrat centered and running along the middle of a furrow in each subplot. Coarse material was defined as litter retained within a 4-mm sieve, whereas fine litter material was defined as litter that was able to pass through a 4-mm sieve.

Soil Textural Components and pH

Soil gravel content was determined by washing, drying, and weighing the gravel collected on a 2-mm sieve from 200 cm−3 of soil from both soil depths within each subplot. Clay, silt, and sand content was determined using particle size analysis [45]. Soil pH was determined on oven-dried (50°C) soil (10 g) in 50 ml of distilled water, shaken for 1 h, and left to stand overnight.

Soil C, N, and P Fractions

Organic carbon (C) was determined by a modified version of the Walkley–Black method and estimated by the equation Ctotal = Corg + Cinorg [49]. Total C and N were determined on oven-dried (50°C) soil using a combustion analyzer (CHN, LECO Corp., USA).

Soil ammonium (\( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \)) and nitrate (\( {\text{NO}}^{{\text{ - }}}_{{\text{3}}} {\text{N}} \)) concentrations were determined using 20 g (oven-dried equivalent) of soil shaken in 80 ml of 0.5 M K2SO4 for 1 h and filtered (Whatman #42). \({\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}}\) and \({\text{NO}}^{{\text{ - }}}_{{\text{3}}} \)–N (\( {\text{NO}}^{{\text{ - }}}_{{\text{3}}} {\text{ + nitrite}} \)) in the K2SO4 extracts were analyzed colorimetrically by automated segmented flow analysis (Skalar Analytical B.V., The Netherlands).

Microbial biomass C (MB-C) was determined by fumigation extraction [4] using 20 g (oven-dried equivalent) of soil shaken in 80 ml of 0.5 M K2SO4 for 1 h. Nonfumigated soils were also extracted. Oxidizable C in the K2SO4 extracts was determined using a Total Oxidizable C analyzer (Simadzu, Japan). A kEC factor of 0.30 was used to calculate the MB-C, as previously determined by Sparling and Zhu [56] for similar Western Australian soils.

Labile inorganic phosphorus (P) and organic P were estimated after extraction of 10 g (oven-dried equivalent) of soil in 50 ml of 0.1 of M NaOH for 16 h [30]. Extractions were immediately filtered, and an aliquot acidified with HCl to precipitate organic matter, refiltered and analyzed for inorganic P using a modified ascorbic acid method [40]. A second aliquot was digested (H2SO4/H2O2) and analyzed for total hydroxide-extractable P, as described above. Labile organic P was estimated as the difference in P between digested and acidified samples.

Community Level Physiological Profiles

A C substrate-utilization method described by Degens and Harris [16] and Degens et al. [17] was used to examine the CLPP of the active microbial populations. In brief, solutions (2 ml) of 25 different organic substrates (as pH adjusted solutions, see below) were each added to 1 g (oven dried equivalent) of moist soil in McCartney bottles (27.7 ml) to form a slurry. Deionized water (2 ml) was added to additional samples to determine respiration response in nonamended soil. All bottles were sealed with a Vacutainer® stopper and incubated at 15°C for 4 h in the dark. After 4 h of incubation, CO2 production was determined by injecting 1 ml of headspace gas into an infrared gas analyzer and analyzed against a known CO2 standard (BOC, Australia). The substrates were the same as those reported in Stevenson et al. [58], consisting of two carbohydrates, two amines, six amino acids, one aromatic acid, and 14 carboxylic acids (listed in Fig. 1). The amines and amino acids were added at 10 mM, the carbohydrates at 75 mM, and the carboxylic acids were added at 100 mM. All solutions were pH adjusted using HCl or NaOH at the time of preparation to individual soil pH to reduce the likelihood of any substrate-pH effects on the microbial populations.

Figure 1
figure 1

Normalized microbial substrate utilization of soils from thinned and burned Jarrah forest. a Mounds (1) 0–5 cm and (2) 5–10 cm soil depths and b furrows (3) 0–5 cm and (4) 5–10 cm soil depths. Error bars represent the standard error of the mean

Full size image

Statistical Design and Analysis

The experimental design was a fully randomized design consisting of four fixed factors: prescribed fire (burned or nonburned), forest thinning (thinned or nonthinned), position (mound or furrow), and soil depth (0–5 cm or 5–10 cm) with three replicates. Differences in soil properties between experimental treatments were tested by fixed effects analysis of variance model, and least significant difference was calculated for the 95% confidence level using Genstat 7.1 (© 2001). Bivariate correlations between soil properties were calculated using Genstat 7.1 (© 2001).

CLPP data were normalized by dividing individual substrate response by total substrate response of each sample, so that differences in total activity did not overwhelm the relative importance of each substrate. Multivariate tests of CLPP data were based on Bray–Curtis dissimilarities calculated among observations without transformation. Tests of the multivariate null hypotheses of no differences among a priori defined groups (four-factor design) were examined using permutation multivariate analysis of variance (PerMANOVA, [1, 43]) and presented in an ordination using unconstrained principal coordinates (PCO, [2]). The F ratio in PerMANOVA is analogous to Fisher’s F ratio and is constructed from sums of squared distances within and between groups. Relationships between CLPP data and soil properties (described in Table 1) were analyzed using nonparametric multivariate multiple regression [43]. Individual variables were analyzed separately for their relationship with multivariate species data (ignoring other variables), and soil properties were then subjected to a step-wise forward selection procedure to develop a model of the species data obtained using 9,999 permutations (DISTLM-forward, [43]). Substrates that “best explained” microbial substrate utilization were based on a subset selection procedure described in detail by Clarke and Warwick [9] and analyzed using the BEST program within Primer v6 software [10].

Table 1 The affect of forest thinning, prescribed fire, sampling position, and soil depth (0–5 cm and 5–10 cm) on soil textural components (%), soil pH, inorganic phosphorus (Pi; mg P kg−1), organic phosphorus (Po; mg P kg−1), organic carbon (OrgC; mg C g−1), total carbon (TotC; %), total nitrogen (TotN; %), soil C/N ratio, coarse litter mass (CL; kg ha−1), fine litter mass (FL; kg ha−1), ammonium concentration (\( {\text{NH}}^{ + }_{4} - {\text{N}} \); mg N kg−1), and microbial biomass C (MB-C; mg C kg−1) n = 3
Full size table

Results

Organic C, total C, total N contents, and soil pH were significantly (P < 0.05) greater in burned compared with nonburned rehabilitated forest, whereas coarse and fine litter mass, silt, sand, and clay contents were significantly (P < 0.05) lower (Table 1). Labile P fractions, MB-C, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), and soil C/N ratio were not affected by prescribed fire (Table 1). Soil gravel content, fine litter mass, soil C/N ratio, and soil pH were significantly (P < 0.05) greater in thinned compared with nonthinned forest, whereas sand and clay contents were significantly (P < 0.05) lower; coarse litter mass, silt, organic C, total C, total N, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), labile P fractions, and MB-C contents were not affected by forest thinning (Table 1). Furrow soil had greater (P < 0.05) coarse and fine litter mass, and labile P, organic C, total C, total N, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), and MB-C contents but lower (P < 0.05) soil pH and soil C/N ratio compared with mound soil; overall, soil textural classes were not affected by contour ripping (Table 1). Soil pH, and labile P, organic C, total C, total N, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), and MB-C contents decreased with soil depth, whereas soil C/N ratio increased (P < 0.05) with soil depth; overall, soil textural classes were not affected by soil depth (Table 1). Soil \({\text{NO}}^{{\text{ - }}}_{{\text{3}}} \)–N was low (<1 μg g−1 soil) in all soils and was not affected by any treatment (data not shown).

Using BEST, 94% of the variation in the CLPP data was associated with the relative utilization of five carboxylic acids (gluconic, malic, l-tartaric, succinic, and uric acids; Fig. 1). The relative utilization of gluconic and malic acids was greater in nonthinned than thinned soil and in mound than furrow soil (Fig. 1). The relative utilization of gluconic and malic acids did not differ between soil depths or between burned and nonburned soil (Fig. 1). The relative utilization of l-tartaric, succinic, and uric acids was greater in thinned than nonthinned soil, mound than furrow soil, and 5–10 cm than 0–5 cm soil (Fig. 1). The relative utilization of l-tartaric, succinic, and uric acids did not differ between burned and nonburned soil (Fig. 1). As a consequence, CLPP were significantly affected by forest thinning (P = 0.035) and contour ripping (P = 0.009) but not by prescribed fire (Fig. 2). CLPP were also significantly affected by soil depth (P = 0.046) but only in nonburned plots (Fig. 2a).

Figure 2
figure 2

Principal coordinates of nonburned (a) and burned (b) plots, which were thinned (T) or nonthinned (NT) and where samples were collected from mounds (M) or furrows (F). Values on Axes 1 and 2 represent the percent of total variance explained by the axes. Solid symbols indicate 0- to 5-cm soil depths, and open symbols indicate 5- to 10-cm soil depths. Error bars represent the standard error of the mean

Full size image

Bivariate correlations revealed that silt, clay, and coarse litter were negatively correlated with soil pH (Table 2). Labile P, organic C, total C, total N, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), and MB-C contents were positively correlated, whereas labile P, organic C and total C, and total N contents were also negatively correlated with soil C/N ratio (Table 2). Total C and total N were also positively correlated with coarse litter mass and negatively correlated with soil C/N ratio (Table 2). \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \) content was positively correlated with MB-C content and negatively correlated with soil C/N ratio (Table 2). \( {\text{NO}}^{{\text{ - }}}_{{\text{3}}} {\text{N}} \) content was positively correlated with fine litter mass (Table 2). The relative utilization of amino acids was positively correlated with organic C and total C contents but negatively correlated with \({\text{NO}}^{{\text{ - }}}_{{\text{3}}} \)–N content and the relative utilization of carboxylic acids (Table 2). The relative utilization of carboxylic acids was positively correlated with \({\text{NO}}^{{\text{ - }}}_{{\text{3}}} \)–N content and negatively correlated with organic C and total C contents (Table 2).

Table 2 Bivariate correlations between selected soil properties
Full size table

Using a multivariate multiple regression model, where variables are introduced individually, all soil properties except labile P contents and fine litter mass explained a significant (P < 0.05) proportion of the variance in CLPP (Table 3). Using a sequential model built with forward selection, the 16 soil variables explained 45% of the variation in CLPP (Table 3).

Table 3 The relationships between soil properties and community-level physiological profiles using nonparametric multivariate multiple regression
Full size table

Discussion

Previous studies have indicated that soil organic matter mediates microbial substrate utilization in arable [18] and grassland [28] soils. Therefore, we proposed that microbial substrate utilization would be correlated with the quantity and quality of soil organic matter present. Indeed, we found using multivariate regressions that coarse litter mass, organic C, total C, total N, \( {\text{NH}}^{{\text{ + }}}_{{\text{4}}} {\text{ - N}} \), \( {\text{NO}}^{{\text{ - }}}_{{\text{3}}} {\text{N}} \), MB-C contents, and soil C/N ratio were all correlated with CLPP in this rehabilitated forest soil (Table 2). These results are consistent with previous studies, which have shown that the chemical nature and diversity of C substrates added to arable [15] or grassland [51] soils can influence substrate utilization. These relationships also support studies that have shown C and N availability mediates microbial functions, including microbial respiration and nutrient mineralisation under Western Australian conditions [12, 29, 30, 47, 48]. Microbial community composition has also been reported to correlate with various C and N pools across a range of ecosystems [3, 11, 12, 35, 42]. Therefore, differences in microbial substrate utilization observed in the current study may be associated with changes in the microbial community composition [31]. However, further examination of soils from rehabilitated forests is required to assess if different management strategies impact on microbial community composition.

Our results indicate that soil textural components varied between burned or thinned and nonburned or nonthinned plots (Table 1). Whereas differences in soil textural components among plots might be attributed to site variation or landscape position; visual observations indicated greater soil erosion in burned and thinned plots than nonburned and nonthinned plots. Whereas forest thinning in our study reduced stem density by 70–80%, the percent plant cover in rehabilitated Jarrah forest has been reported to be reduced from 30–50% to less than 10% after a prescribed burn [23, 25]. Therefore, it is possible that prescribed fire and forest thinning allowed greater quantities of rain to fall directly onto the forest floor, carrying silt, clay, and sand components either down the soil profile or down slope. It should also be noted that two thinned plots were also situated on a relatively steep slope (15°–20°) where soil erosion was more pronounced. Subsequently, the influence of forest thinning on microbial substrate utilization was also considerably stronger in these two plots, evident by the large lateral error bars associated with PCO analysis of CLPP from forest thinned soil (Fig. 1). However, soil samples were not collected before the commencement of our study, so differences in textural components may have occurred previous to the treatments being implemented. Therefore, differences in soil erosion may have been wholly caused by the landscape position, although variability in the textural components of the initial topsoil placed in the mined pit can also not be discounted. It is also possible that such differences in initial soil textural components may have been exacerbated by prescribed fire and forest thinning management in our study.

Soil texture is considered a primary mediating factor of microbial function via its influence on the retention of organic matter and water [33, 68]. However, soil textural components and organic matter fractions were not correlated in our study (Table 2). Soil texture may also mediate the number of soil niche environments and the ability of the microbial population to protect itself from desiccation [5] and predators [53]. Previous studies have also reported concurrent changes in substrate utilization and microbial community composition [51]. Therefore, we propose that the significant relationships between soil textural classes and CLPP revealed by multivariate regression in our study are likely to be associated with the influence of soil textural components on microbial community composition. These results demonstrate that relatively small changes in soil textural components can have considerable effect on microbial substrate utilization in this mined Jarrah forest ecosystem.

Forest thinning and contour ripping has been shown to mediate litter, organic matter, and nutrient content of surface soil in both the current study (Table 1) and elsewhere [46, 50, 62]. Therefore, we proposed that forest thinning and contour ripping would directly influence microbial substrate utilization. We found that substrate utilization differed between thinned and nonthinned soil and also between mound and furrow soil (Fig. 2). These results were largely (94%) produced by differences in the relative utilization of gluconic, malic tartaric, succinic, and uric acids associated with forest thinning, contour ripping, and soil depth (Fig. 1). Low molecular weight carboxylic acids such as these are probably intermediate products of microbial metabolism, leachate from forest litter, or root exudates [19, 34, 60]. Therefore, differences in soil textural classes, soil pH, litter mass, and C and N pools induced by forest thinning and contour ripping are likely to influence the availability of these organic acids to the microbial community [19, 37, 54, 59, 61] in our study. Subsequently, the availability of low molecular weight organic acids can change both substrate utilization and microbial community composition [51]. However, low molecular weight organic acids are generally present in forest soils at concentrations of 0–1 mM [64–66], which is many orders of magnitude less than the concentrations applied in the CLPP technique (10–100 mM). These organic acids are readily assimilated by various heterotrophic microorganisms and will occur in soils under dynamic equilibrium conditions between their production and consumption/absorption processes [60]. Therefore, differences in the relative utilization of these organic acids most likely reflect differences in microbial nutrient demand [15, 51], which, in our study, will result from management induced differences in soil properties. Similarly, a recent study reported that the addition of complex compounds such as starch to a grassland soil produced a 10% greater proportional CLPP utilization of N-containing amino acids than did a simple compound such as glucose [51]. Orwin et al. [51] concluded that more enzymes were required for the decomposition of more complex C compounds, which resulted in greater microbial N demand. Similarly, we found that the relative utilization of amino acids tended to increase with increasing total C and organic C contents but decreased with increasing \({\text{NO}}^{{\text{ - }}}_{{\text{3}}} \)–N content, the opposite was true for carboxylic acids (Table 2).

Forest productivity and soil nutrient status has been shown to recover within 4–6 years of prescribed fire in rehabilitated Jarrah forest [46]. Other studies have also shown that 5 years after burning, BIOLOGTM substrate utilization was the same in prescribed burned and nonburned pine forest soils [57]. In our study, 1 year after a spring prescribed fire, litter plus C, N, and P pools were either not different or greater in burned than nonburned forest soils (Table 1). These results are likely to have resulted from litter reaccumulation being rapid (2 weeks) after a prescribed burn in rehabilitated Jarrah forest, with rates two to three times that of nonburned sites [55] and spring burns generally being low intensity [25]. Previous studies have also shown that burning rehabilitation sites stimulates the growth of N-fixing legumes [25]; however, we found no evidence of this 1 year after a prescribed burn. Despite these differences in litter and organic matter components, there was no difference in the relative utilization of CLPP substrates between burned and nonburned soils (Fig. 1). This lack of difference suggests that fire induced changes to the quantity of organic matter has not influenced CLPP in this forest soil. We conclude that CLPP in this mined Jarrah forest ecosystem is resilient to a spring prescribed fire and that any possible fire effect occurs during the first year after fire. However, recent studies have also proposed that improved ecosystem functional information can be gained from CLPP if substrate selection is optimized to better reflect more relevant changes in the soil environment to be tested [6, 41]. For example, application of substrates at lower concentrations that better reflect natural soil conditions is possible using 14C-labeling, which would enable detection of CO2 response at much lower rates of activity [6]. Furthermore, optimization of the CLPP method such that the effect of specific fire induced changes to soil organic matter composition (i.e. (1) removal of external oxygen groups; (2) reduction in chain length of alkyl compounds, such as alkanes, fatty acids, and alcohols; (3) aromatization of sugars and lipids; (4) formation of heterocyclic N compounds; (5) condensation of humic substances; and (6) production of an almost unalterable component, black carbon [7]) on substrate utilization can be tested requires further examination [41].

In both our studies and elsewhere, contour ripping and soil depth have been reported as the major sources of organic matter and nutrient heterogeneity in rehabilitated Jarrah forest soils [62, 67]. Our study indicates that microbial substrate utilization was also affected by contour ripping and to a lesser extent by soil depth. We proposed that prescribed fire would decrease the contour ripping and soil depth induced organic matter heterogeneity via the combustion of surface litter. We also proposed that forest thinning would increase organic matter heterogeneity as in the short-term, thinned organic material would accumulate in surface soil and preferentially within furrows. However, we found no significant interactions between prescribed fire or forest thinning and contour ripping and soil depth when we analyzed any soil properties or CLPP data. These results suggest that the effect of prescribed fire and forest thinning on organic matter, nutrient availability, or CLPP in our study did not influence the spatial heterogeneity that already existed in nontreated rehabilitated Jarrah forest soil.

As described above, we found correlations among soil textural classes, litter mass, soil C and N pools, and CLPP, which agrees with similar relationships reported elsewhere [18, 22, 28]. Although these relationships improve our understanding of those properties that mediate microbial functioning, characterizing the relative importance of these soil properties has received little attention. When covariance between soil properties was accounted for, only half of the variance in CLPP data was explained in our study (Table 3). The CLPP technique used in our study measures the respiration response from the addition of simple C substrates to whole soil [6, 16]. The availability, composition, and diversity of similar C compounds as used in the CLPP technique have been shown previously to mediate the activity of that portion of the microbial community responsible for the CLPP respiration [15, 51]. Such compounds are often dissolved and mobile within the soil solution and are thus generally considered to have a major role in the transport and supply of C and N to microbial populations [13, 14, 36, 38]. Therefore, the characterization of the nutrient components within the dissolved organic matter pools may provide improved explanation of CLPP variance among soils in future studies.

In conclusion, CLPP were not affected by low-intensity prescribed fire, although soil properties that were correlated with CLPP were affected. CLPP were affected by forest thinning, which corresponded to changes in litter mass, organic matter quality, and soil pH. Differences in soil textural components between treatments were also correlated with CLPP. Reasons for differences in soil textural components in our study include landscape position, and/or postmining rehabilitation management induced variation in soil erosion or initial textural components. Our results suggest that 1 year after treatment, CLPP from this mined forest ecosystem are resilient to a spring prescribed (low intensity) fire but not forest thinning. We conclude that differences in CLPP are likely to result from complex interactions among soil properties that mediate substrate availability, microbial nutrient demand, and microbial community composition.