FLAMMA, 4 (1), 41-45, 2013 ISSN 2171 - 665X CC Attribution-NonCommercial-ShareAlike 3.0 License

Soil microbial activity in Aleppo pine stands naturally regenerated after fire: silvicultural management and induced drought D. Moya (*), J. Hedo, M. Andrés, C. Wic, J. de las Heras, F.R. López-Serrano, E. Rubio Universidad de Castilla-La Mancha, Campus Universitario s/n, 02071, Albacete, Spain * Corresponding author: [email protected]

Keywords

Abstract

Forest management Fire ecology Induced drought Pinus halepensis

In post-fire restoration, early monitoring is mandatory to check impacts and ecosystem responses to apply proper management according to social standards and ecological conditions. In areas where the natural regeneration was successful, excessive tree density can be found which induces to high intraspecific competence and assisted restoration management could be adequate. In addition, climatic changes will have large impacts on vegetation productivity and resilience since the regional models for south-eastern Spain predicts a rainfall decrease of about 20% and temperature o increase of 4.5 C. The microbial biomass could be used as indicator of ecosystem recovery, since it is negatively affected by wildfires and depends on fire characteristics, vegetation and soil properties. Our aim is to determine how forest management may affect the ecosystem recovery in different climatic scenarios, included drought scenarios with and without forest management (thinning). We compared soil physicochemical properties and microbial activity in four scenarios: unmanaged and thinned stands in two rainfall scenarios (under induced drought). The study areas were set close to Yeste (Albacete) where Aleppo 2 pine forest were burned in summer 1994 (nearly 14000 ha). We set sixteen rectangular plots (150 m ; 15 m ×10 m) implementing experimental silvicultural treatments: thinning eight plots in 2004, reducing the naturally recovered tree -1 density from about 12000 to 1600 pine trees ha . In addition, in half the plots, we induced drought conditions from about 500 to 400 mm (20%) from March 2009. In every plot, we monitored temperature at ground level (Ts), 10 cm depth (T10d) and soil relative humidity (RH). Taking into account season of the year and canopy coverage, we collected soil samples in mid-winter (ending January 2011) and mid-spring (ending May 2011) under pine trees and in bare soil. The soil samples were used to evaluate soil physicochemical properties and soil microbial activity. The latter was indirectly analysed and was characterized by measuring a set of enzymatic activities (urease, phosphatase, bglucosidase and dehydrogenase). Neither significant differences in the values of physicochemical properties (with exception of soil organic matter) nor in temperatures were found. Regarding the microbial activity, we found differences on the significance of factors depending on the enzymes. The results obtained indicate that ecosystem response can be modified by silvicultural treatments, at least in the short term. The soil microbial data can be evaluated using enzyme activity which should be integrated in the vegetation pattern monitoring to improve vulnerability assessments, mainly in areas prone to wildfires and to suffer climate change. The proposed silvicultural treatments could be applied as a rehabilitation tool to enrich soil and vegetation, increasing soil resource availability, restoring microbial properties at short term, and improving the resilience of the ecosystem.

1

change models foresee changes in droughts and fire regimes (Flannigan et al., 2000). The effect of forest fires on the ecosystem depends on several factors, mainly related to fire severity, structural patterns, weather and

INTRODUCTION

Forest fires are important in landscape modelling along the Mediterranean Basin, but predictions based on climate 41

FLAMMA | Vol. 4 | 1 | 41-45

landform (Ryan et al., 2005). In post-fire restoration, early monitoring is mandatory to check impacts and ecosystem responses, applying management strategies according to social standards and ecological conditions to recover and improve the ecosystem and their resilience (De las Heras et al., 2012). The microbial biomass is negatively affected by wildfires, which reduce the microbiological activity and their metabolic activity, depending on fire characteristics, vegetation and soil properties (Guenon et al., 2011).

2

Juniperus oxycedrus L., Rosmarinus officinalis L. and Quercus coccifera L.

3.2 EXPERIMENTAL DESIGN 2

We set sixteen rectangular 15 × 10 m plots in the burned and recovered area, implementing experimental silvicultural treatments, in early 2004 we thinned eight plots, reducing the natural tree density from about 12000 -1 to 1600 pine trees ha . In half the plots, we simulated an increase in drought conditions; rainfall was artificially reduced from about 500 to 400 mm. The final experimental design was two bivariate factors: silvicultural treatment (TREE DENSITY) and induced drought (RAINFALL). Finally, we had 4 control plots (high tree density), 4 control+drought (high tree density and low rainfall), 4 thinned (low tree density) and 4 thinned+drought (low tree density and low rainfall).

OBJECTIVES

Our main objective is to determine how forest management may affect the ecosystem recovery. The study was focused in Aleppo pine stands naturally recovered after a fire which were thinned. We included drought simulation to check effects in different climatic scenarios. This is a short-term study, since sampling tasks were carried out three years after the treatments. To characterize the ecosystem response after fire and the assisted restoration effect in two rainfall scenarios, we compared soil physicochemical properties and microbial activity in managed and unmanaged stands, including those under induced drought.

3

In the plots under simulated drought, water availability was restricted since March 2009 until the end of the study (June 2011) by installing collectors (PVC half-pipes), suspended about 30 cm above the soil surface, and digging trenches, (0.25 m in depth), on the upper sides of the plots to prevent lateral water flow. The plastic structures covered about 20% of the surface area of the plots (draining about 20% of rainfall). We monitored temperature at ground level (Ts), 10 cm depth (T10d) and soil relative humidity (RH).

METHODOLOGY

3.1 STUDY SITE In summer 1994, a large forest fire burned nearly 14000 ha in SE Spain, close to Yeste (Albacete). Before the fire, vegetation consisted of a mixed scrubland made up by Pinus halepensis Mill. and P. pinaster Aiton. Average rainfall and temperature values from 1986 to 2011 (based upon data provided by the Spanish National Meteorological Agency, AEMET), were 495 mm and 14.8 o C, respectively. The ombroclimate was characterized as Dry, located in the Mesomediterranean belt (Rivas Martinez, 1982). According to the Spanish Soil Map (IGN, 2006) and Soil Taxonomy (Soil Survey Staff, 2010), soils in the area were classified as Xerorthents. In general, soils in the area are shallow (thickness between 20-50 cm), coarse-textured (sandy), and alkaline (soil pH>8). The experiment was conducted on nearly flat areas (slope ≤ 2%) with naturally regenerated P. halepensis stands. The natural potential vegetation was a sclerophyllous oak forest of Quercus ilex (Bupleuro rigidi-Querceto rotundifoliae sigmetum) (Rivas Martinez, 1982). After fire, the area was naturally regenerated predominating P. halepensis (as main tree species), Thymus vulgaris L.,

3.3 SAMPLE COLLECTION AND ANALYSIS The sampling collection was executed in two different seasons of the year (SEASON): mid-winter (ending January 2011) and mid-spring (ending May 2011) to check differences on soil microbial activity depending on photosynthetic activity (Prieto et al., 2009). Taking into account the canopy coverage (COVER), we took soil samples under pine trees and in bare soil (Moya et al., 2009). The soil samples were composed of three subsamples collected at each plot, stored in sealed plastic bags and carried to the laboratory. Soil physicochemical properties (pH, electrical conductivity, soil organic matter (SOM), soil respiration and total organic carbon (TOC)) -1 were recorded. Microbial activity (unit g ) was characterized by measuring a set of enzymatic activities, namely urease (Kandeler et al., 1999), phosphatase and bglucosidase (Tabatai & Bremner, 1969) and dehydrogenase (Garcia et al., 1997).

42

FLAMMA | Vol. 4 | 1 | 41-45

A

Figure 2. Estimated response surface analysis using the most influential factors obtained in a General Linear Model for dehydrogenase activity (upper image) and -glucosidase activity sampled in winter (lower-left image) and spring (lower-right -1 image). Units: g .

B

(Least Significant Difference) method. The values were transformed using logarithms to maximize the Pearson’s correlation coefficient (r), achieving normality assumptions and homoscedasticity. However, the values are shown untransformed along with standard error (±SE). Statistical analyses were carried out using Statgraphics CENTURION XV software and conducted using a critical p-value<0.05.

4

RESULTS AND CONCLUSIONS

Neither significant differences in the values of physicochemical properties (with exception of soil organic matter) nor in in temperatures (Ts and T10d) were found. The GLM analysis pointed to relationships of SOM (R = 2 69.71%) and RH (R = 68.67%) to the factors TREE DENSITY and RAINFALL, however the main factor influencing SOM was COVER, while for RH was SEASON. ANOVAs showed higher RH in mid-winter (22.68 ± 0.54%) than in mid-spring (5.54 ± 0.29%). Comparing within seasonal values, we found lower values in plots with low tree density and low rainfall. Regarding SOM, ANOVAs showed higher values in covered areas (8.45±0.14%) than in bare soils (7.99 ± 0.12%). The tendency varied depending on COVER: bare soils showed higher values in thinned plots and lower in control plus induced drought but in those under pine tree coverage, the highest SOM was found in control (no differences were found in other plots).

Figure 1. Estimated response surface analysis using the most influential factors obtained in a General Linear Model for urease -1 (left) and phosphatase (right) activity (units: g ) obtained in two different seasons (A, winter, and B, spring) and two different covers (under pine coverage (right images) and bare soil (left images).

2

3.4 STATISTICAL ANALYSIS General Linear Models (GLM) were applied to assess an overall model to select factors (SEASON, COVER, TREE DENSITY and RAINFALL) and their interactions which contributed significantly to predict variables (according higher values for F-ratio). We used Kruskal-Wallis tests to check mean significant differences among independent variables. Significant differences between groups were tested using simple ANOVA developed with Fisher's LSD 43

FLAMMA | Vol. 4 | 1 | 41-45

Regarding the microbial activity characterized by measuring enzymatic activities, GLM showed models which included significant factors (and interactions) 2 explaining high variability for urease (R = 87.19%), 2 dehydrogenase (R = 71.07%) and phosphatase (R = 2 49.749%) but low for b-glucosidase (R = 28.37%). The main influential factors for the four selected variables were:

microbial properties at short term, and improving the resilience of the ecosystem.

ACKNOWLEDGEMENTS

2

We would thank to the Regional Forestry Service of Castilla-La Mancha for their help and support. Also to the projects of the I+D+I National Programme which provided funds: FIREMED3L: AGL2011-27747 and CONSOLIDERINGENIO 2010: MONTES (CSD 2008-00040). In addition, special thanks are given to the COST Office, the EU Framework Programme and European Science Foundation for the "Conference Grant for Early Stage Researchers" which allowed to Daniel Moya assist the meeting (COSTCONG-DC_FPS-00111).



UREASE (Figure 1): the two main interactions were SEASON*RAINFALL (F-ratio=42.38) and SEASON*CUBIERTA*TREE DENSITY*RAINFALL (Fratio=11.20).  PHOSPHATASE (Figure 1): SEASON (27.83) and interaction COVER*TREE DENSITY*RAINFALL (6.00) were the most influential.  DEHYDROGENASE (Figure 2): TREE DENSITY and TREE DENSITY*RAINFALL were showing the higher F-ratios (23.93 and 24.72, respectively).  ß-GLUCOSIDASE (Figure 2): the most influential interactions were SEASON*TREE DENSITY (Fratio=11.73) and SEASON*RAINFALL (Fratio=8.20). In general, urease activity is positively linked to vegetable activity (spatially and periodically), being sensitive and negatively linked to drought. Phosphatase activity decreased in spring but increased in drought plots according to lower tree density. Dehydrogenase was neither linked to coverage nor season of the year and was showed higher values in both thinned plots (normal rainfall) and control plus induced drought. Finally, bglucosidase activity depended on the season of the year, being higher in winter according lower tree density but in spring, we found the highest value in control plots. So, thinning was not affecting urease activity, increased phosphatase (mainly under induced drought), promoted dehydrogenase activity (normal rainfall) and affected bglucosidase depending on the season of the year.

REFERENCES Bates BC, Kundzewicz ZW, Wu S, Palutikof J.P. 2008. Climate Change and Water. Technical Paper of the Intergovernmental Panel on Climate Change, IPCC Secretariat, Geneva. De las Heras J, Moya D, Vega JA, Daskalakou E, Vallejo R, Grigoriadis N, Tsitsoni T, Baeza J, Valdecantos A, Fernández C. 2012. Post-Fire Management of Serotinous Pine Forests. In: Moreira F, Arianoutsou M, Corona P, De las Heras J. (Eds.), Post-Fire Management and Restoration of Southern European Forests. Managing Forest Ecosystems 24, pp. 121–150. Flannigan MD, Stocks BJ, Wotton BM. 2000. Climate change and forest fires. Science of the Total Environment 262: 221–229. Garcia C, Hernandez T, Costa F. 1997. Potential use of dehydrogenase activity as an index of microbial activity in degraded soils. Communications in Soil Science and Plant Analysis 28: 123–134. Guenon R, Vennetier M, Pailler A, Dupuy N, Roussos S, Gros R. 2011. Trends in recovery of Mediterranean soil chemical properties and microbial activities after infrequent and frequent wildfires. Land Degradation and Development. DOI: 10.1002/ldr.1109. IGN. 2006. Mapa de suelos de España. Instituto Geográfico Nacional, Madrid, Spain. Kandeler E, Stemmer M, Klimanek EM. 1999. Response of soil microbial biomass, urease and xylanase within particle size fractions to long-term soil management. Soil Biology and Biochemistry 31: 261–273. Moya D, Hedo J, Andrés M, Wic C, de las Heras J, LópezSerrano FR, Ferrandis P, Rubio E, Lucas-Borja ME. 2009. Soil response, microbial activity and natural regeneration after fire passage in a semiarid

Aleppo pine forests have developed adaptions to fire and drought events, but they could be not prepared for the scenarios predicted by climate change models (Bates et al., 2008). The results obtained indicate that ecosystem response can be modified by silvicultural treatments at short term. The soil microbial data can be evaluated using enzyme activity, by using urease, dehydrogenase, phosphatase and b-glucosidase, which should be integrated in the vegetation pattern monitoring, in order to improve vulnerability assessments under frequent wildfires and climate change in Mediterranean ecosystems. In addition, the silvicultural treatments could be applied as a rehabilitation tool to enrich soil and vegetation, increasing soil resource availability, restoring 44

FLAMMA | Vol. 4 | 1 | 41-45

Aleppo pine forest. In: Jordán A, Zavala LM, de la Rosa JM, Knicker H, González-Pérez JA, GonzálezVila FJ. (eds). Book of Abstracts of the Communications presented to the II International Meeting on Forest Fire Effects on soils, FUEGORED 2009, Sevilla – Cortegana, Spain, pp. 25–26. Prieto P, Peñuelas J, Llusià J, Asensio D, Estiarte M. 2009. Effects of experimental warming and drought on biomass accumulation in a Mediterranean shrubland. Plant Ecology 205: 179–191. Rivas Martinez S. 1982. Etages bioclimatiques, secteurs chorologiques et série de végétation de l'Espagne méditerranéenne. Ecologia mediterranea 8: 275– 288. Ryan KC, Debano LF. 2005. Wildland fire in ecosystems: effects of fire on soil and water. Gen. Tech. Rep. rmrs-gtr-42 (4) U.S. Department of Agriculture, Forest service, Rocky Mountain research station, pp. 250. Soil Survey Staff. 2010. Keys to Soil Taxonomy, 11th ed. USDA-Natural Resources Conservation Service, Washington, DC. Tabatabai, MA, Bremner JM. 1969. Use of p-nitrophenyl phosphate for assay of soil phosphatase activity. Soil Biology and Biochemistry 1: 301–307.

45