Quantifying the post-fire carbon exchange in a spanish black pine (pinus nigra arn. Ssp. Salzmannii) forestestimates based on biometric and eddy-covariance methods

Resumen

Forest ecosystems are a vital part of the global carbon (C) cycle. Their potential to act as natural sinks of atmospheric CO2 is well known, being able to slow or reverse the increasing concentration of atmospheric CO2, which can contribute to reduce global warming and to prevent climate change. Therefore, it is necessary to deeply understand the ecosystem-atmosphere C exchange from the different forest biomes worldwide. Nevertheless, besides their importance in the Mediterranean basin, the C sink strength of Spanish black pine (Pinus nigra Arn. spp. salzmannii) forest ecosystems under natural conditions yet remains unknown. Besides, the current and future trends in Mediterranean fire regime as well as the null or limited post-fire resilience of this pine species make it a critical issue requiring additional efforts to address the lack of information regarding the forest-atmosphere interaction in recently burnt Spanish black pine stands under different degrees of burn severity. As well, it is worth to note that unravelling how slope-aspect modulates post-fire C fluxes within severely burnt landscape has yet to be addressed in this fire-disturbed pine forest ecosystems. In order to better understand the significance of different forest ecosystem C cycling processes and to increase the reliability and consistency of the estimates of the net ecosystem C exchange and its main component fluxes, a cross-validation of these C fluxes estimated by fully independent approaches with unrelated errors should be convenient in C cycling studies. Currently, the biometric and flux chamber-based methods (BM) and the eddy-covariance technique (EC) are the foremost methodologies to quantify the ecosystem-atmosphere C exchange, although the analysis of the consistency of BM- and EC-based C fluxes is fairly difficult; therefore, only few studies have conducted comparisons at both site-level or global-scale. Within this framework, this Thesis examines over 3 years (2011-2013) the magnitude, temporal and spatial patterns, and environmental controls of the net ecosystem C exchange and its component C fluxes in a salvage-logged Spanish black pine forest located at the Cuenca Mountain Range Natural Park (Castilla-La Mancha Region, central-eastern Spain). We measured C fluxes along a burn severity gradient (unburnt, low burn-severity, and high burn-severity). Concretely, this work focuses on early stages following fire (1.5-4.5 years post-burn) comprising simultaneous measurements of BM and EC approaches. In addition, this Thesis also assess the spatio-temporal patterns of post-fire C fluxes from both the soil and decaying tree stumps between opposing slope-aspects (north- vs. south-facing) within the severely burnt landscape. Thus, after conventional salvage logging was performed, four experimental sites (approx. two ha each, about 500 m apart) were established based on a burn-severity map: i) an unburnt control site (UB, mature forest), ii) a low burn-severity site (LS), iii) a south-facing high burn-severity site (HSS), and iv) a north-facing high burn-severity site (HSN). The experiments included in the Chapters of this Thesis monitored and quantified, at ecosystem-level, the main C fluxes at unburnt and burnt sites by using different methodologies. The C loss from decaying tree stumps was estimated via wood mass loss (indirect method) and via in situ chamber-based CO2 flux measurements (direct method) at the LS, HSS, and HSN sites (Chapter 1). The total soil respiration was measured through both automated and manual chamber-based measurements along the burn severity gradient (UB, LS, HSS, and HSN sites; Chapter 2). The aboveground autotrophic respiration was determined at both UB and LS sites through the analysis of complex and simple scale-up methods. These methods did or did not account for, respectively: i) the vertical variation in the total wood CO2 efflux within individual pine trees, and ii) the effects of both the leaf ageing (current- vs. previous-years needles) and light inhibition (darkness vs. light) on foliar respiration (Chapter 3). The gross primary production (GPP) was estimated at both the UB and LS sites through two different modelling approaches. One of them was based on the C-mass balance, which was calculated as the sum of the aboveground net primary production, the total belowground C flux, and the aboveground autotrophic respiration. The other approach was based on the whole-canopy photosynthesis-modelling, which was obtained by combining an environmental-dependent non-rectangular hyperbolic light-response model, applied to different pine needle age-cohorts, and coupled to a two-leaf scaling-up strategy (i.e., sunlit and shaded canopy fractions). These GPPC and GPPM estimates, respectively, were then used in order to estimate the net ecosystem production (NEPC and NEPM, respectively) at both sites from ecosystem respiration (Reco) estimates obtained in previous Chapters (Chapter 4). Finally, the net ecosystem C exchange (NEE or NEPEC) and its components (GPP, and Reco) were assessed at both the LS and HSS sites by using year-round (January-December 2012) eddy-covariance (EC) measurements (Chapter 5). Our findings revealed that the UB site acted as moderate C sink over the whole study period (2011-2013), with a mean annual NEPC and NEPM of 2.43 and 2.04 Mg C ha-1 year-1, respectively. This study ascertained that the GPP was the largest C flux at this mature pine forest. The mean relative contribution of soil, aboveground wood tissues (stem+branches), and needles to the Reco was 51, 26 and 23%, respectively, confirming that the total soil respiration was the major source of C release involved in the ecosystem-atmosphere C exchange at this site. This research showed that the low severity surface fire did not substantially alter the GPP, Reco, and NEP estimates compared to the nearby UB site; therefore, the LS site also acted as a moderate C sink (i.e., the mean annual NEPC and NEPM estimates over the period 2011-2013 were 2.09 and 1.82 Mg C ha-1 year-1, respectively; the annual NEPEC estimate in 2012 was 2.69 Mg C ha-1 year-1). However, the high severity stand-replacing fire, which shifted the forest ecosystem into a sparse grassland, induced a strong alteration on the ecosystem C exchange that lead to the HSS site behave as a moderate C source (i.e., the annual NEPEC estimate in 2012 was -2.08 Mg C ha-1 year-1). Tree stumps can be considered as hot spots of CO2 production during their early stages of decay, although their contribution to the Reco, which was significantly higher at both the HSS and HSN sites than those at the LS site, was insignificant at this burnt forest ecosystem. Low severity surface fire induced a limited alteration of the total soil respiration rates compared to the UB site because similar abiotic and biotic factors were observed at both unburnt and low burn-severity sites. In contrast, both the HSS and HSN sites showed a general increase in total soil respiration rates than those at the UB site, which was attributed to the influence of ameliorating soil environmental conditions, but especially to the C flux from decaying stump roots. Overall, our diachronic experiment has also revealed that more C was released from both the soil and decaying tree stumps at the south-facing slope-aspect, which occurred under warmer and drier conditions than those occurred at the north-facing slope aspect. In addition, our findings showed that the low severity surface fire did not induce the modification of the aboveground (wood+needles) respiratory processes in the remaining pine trees after fire compared to the unburnt ones. Based on simultaneous BM and EC measurements performed at the LS site during 2012, our findings showed that ecosystem-level observations obtained by the EC approach provided analogous estimates of annual GPP but slightly smaller estimates of annual Reco compared to those from BM approach. Consequently, it resulted in a relatively low overestimation of annual EC-based NEP estimates in comparison with those obtained by BM approach. The discrepancy between both EC- and BM-based Reco estimates was associated with complex error sources. Overall, the high convergence between BM-EC C fluxes evidenced the reliability and consistency of both approaches to quantify the ecosystem-atmosphere C exchange at this burnt pine forest ecosystem. In this study, we also built on efforts to verify common sampling methodologies as well as to develop innovative strategies to obtain accurate BM-based C flux estimates. Our work confirmed that the C flux from partially buried stumps left on site was more accurately characterized by the direct (CO2 flux chamber) method rather than the indirect (wood mass loss) method. Our findings established that a combined effort with automated and manual chamber-based measurements helped to enhance the accuracy of upscaled total soil respiration rates by reducing bias associated with the non-optimally timed manual sampling. This study also suggested that the quantitative analysis of wood CO2 efflux regarding both the tree size distribution within the stand as well as the vertical variation within individual trees allowed providing improved ecosystem-level estimates. Our study highlighted the need to incorporate both the needle age and light inhibition variations in the needle respiration when estimating this C flux at the ecosystem-level. Although several simplifying assumptions were implicated, the 3-year convergence of annual GPP estimates obtained by both the C-mass balance and whole-canopy photosynthesis-modelling approaches served as an important cross-validation, demonstrating that these laborious and independent approaches provided reliable ecosystem-level GPP estimates. Overall, this Thesis represents a considerable step forward to fill gaps in knowledge concerning the ecosystem-atmosphere C exchange in burnt Spanish black pine stands during the early stages following fire. This study also contributes to better understanding of the methodologies for monitoring and quantifying ecosystem C exchange through the simultaneous application of both the BM and EC approaches.