Resilience against fire disturbance of Mediterranean vegetation has been frequently described. However, due to climatic change and abandonment of local land use practices, the fire regime is changing, probably leading to higher intensities and frequencies of disturbance events. The forthcoming scenario calls for a full understanding of post-disturbance tree recruitment processes, structural resilience and possible consequences on the overall forest biodiversity. In particular, knowledge on severe crown fires’ effects on forest stand structural attributes needs to be further explored. In this work, we describe and quantify fire impact and short-term response of a Mediterranean forest affected by high severity crown fires, focusing on the compositional and structural diversity of living and dead trees, spatial pattern of fire-induced mortality, recovery dynamics of tree species. The analysis, based on a synchronic approach, was carried out within four burned and two not burned fully stem-mapped research plots located in NW Italy, belonging to two forest categories differing for their main tree restoration strategies. Distance-dependent and distance-independent indices were applied to assess structural diversity dynamics over time since fire occurrence. Within the analyzed forests fire was found to affect mostly forest structure rather than its composition. Number of snags largely increases immediately after the fire, but it levels off due to their fall dynamics. Regeneration strategies and fire severity influenced species abundance and consequently diversity patterns. Stem diameter and height diversity were modified as well, with a strong increase in the first post-fire year and a sharp reduction six years after the disturbance. Fire determined also a higher heterogeneity in crown cover and vertical structure. Spatial patterns of surviving trees and snags were greatly affected by fire, producing an increase in aggregation and segregation mechanisms. Autosuccessional processes are supposed to preserve current forest structure and composition, but the ecosystem self-restoring capability should be analyzed in the light of the possible changes in fire regime.
Fire is recognized as the most important natural disturbance in Mediterranean ecosystems (
Even though fires have been a major factor in Mediterranean ecosystems for millennia, the general trend in number of fires and area burned in European Mediterranean areas has dramatically increased during the last decades, principally due to land-use and climatic changes (
Although Mediterranean vegetation is able to cope with fire (
It is well documented that Mediterranean ecosystems recover readily after fire through an autosuccessional process (
The main aim of this work was to describe and quantify short-term vegetation response after high-severity crown fires in a Mediterranean forest with stands featuring a different composition of seeder and resprouter woody species in the tree layer. Recovery processes were investigated to disclose differences in resilience mechanisms among species with different restoration strategies. Specific research goals were the assessment of: (a) compositional and structural diversity of living and dead trees; (b) spatial pattern of fire-induced mortality; (c) recovery dynamics of tree species.
The study was carried out in the province of Savona, Central Liguria, NW Italy (lat 44° 12’ - 44° 19’ N, long 8° 22’ - 8° 28’ E). The geologic bedrock is quite heterogeneous and is mainly represented by Cenozoic and Mesozoic igneous and metamorphic rocks. The climate is typical Mediterranean, with mild and moderately moist winters and warm and dry summers, characterized by a significant summer drought. Along with a few remaining monospecific
Two high severity crown fire events, located in close proximity (< 10 km), were selected for a case study. The first wildfire occurred in August 1998 and burned 146 ha, while the second occurred in August 2003 and burned 235 ha. Burned areas within the fire perimeters were differentiated according to the percentage of seeder and sprouter species in the overstory, identifying two main forest categories: the mixed-broadleaved forest category (henceforth: MB) characterized by the predominance of sprouters, and the broadleaved-coniferous forest category (BC) made up by both sprouters and obligate seeders.
We established one intensive sampling plot for each forest category inside the fire perimeters (BC_03; BC_98; MB_03; MB_98) and another one for each forest category in the surrounding unburned area as control plot (BC_nb; MB_nb). The latter were not burned by high intensity fires since 20 years at least.
Each plot (50 x 50 m) has one side parallel to the maximum slope. Within each plot all living trees and shrubs (stems with dbh ≥ 3 cm), snags (standing dead trees with dbh ≥ 3 cm and height > 1 m), logs (downed dead trees or part of, with the biggest size diameter ≥ 10 cm), and stumps (surface diameter ≥ 10 cm and height ≤ 1 m) were identified, labeled with numbered plastic tags and mapped with a total survey station (Geotronics - Geodimeter 400). Sampling activities were carried out in summer 2004.
For every living stem, we recorded species, diameter at breast height (dbh), total height, height of lowest living branches (upslope and downslope), four radii of the crown projection on the ground. For each snag, log or stump inside the plot, the species based on distinguishing traits not modified by fire, the diameter (dbh for snags, diameter on both ends for logs, diameter at the root collar for stumps) and the total height (snags, stumps) or length (logs) were recorded. The number of post-fire resprouts of tree and shrub species was also recorded. Tree seedlings and saplings were recorded within sixteen square subplots (6.25 m2) randomly established inside each 50 x 50 m plot.
Structural diversity within plots was assessed applying both distance-independent and distance-dependent measures. For more details on the diversity indices used, see
The Brillouin index (
Following
The vertical distribution of canopy cover within the plots was assessed through the Vertical Evenness (VE) index by
To describe the tree spatial pattern within the plots, Point Pattern Analysis (PPA) techniques were applied by means of Ripley’s
To determine the distributions of trees or tree-classes as random, regular or clumped, univariate Ripley’s
We considered as living the individual stools showing at least one shoot (dbh ≥ 3 cm) alive. All the analyses were carried out only for classes having more than 20 elements, and starting from 1 m up to 25 m applying a 1 m lag distance, thus not exceeding half of the study area in order to limit the influence of the margin effects (
Null models were chosen for the different analyses to avoid misinterpretation of the results (
A total of 6482 living stems, snags, logs and stumps were recorded, mapped and measured in the tree layer of the 6 investigated plots. Living stem density (
Comparable values in density of seed germinated individuals were found for unburned plots and those burned in 2003 (Mann-Whitney test; p>0.05), whereas a highly significant difference occurred in 1998-burned plots (Mann-Whitney test; p<0.001)
The size-class distribution of living stems and snags was reverse
Among the most abundant species in the overstory,
Higher species diversity values (
Living stems of both categories had similar THD and TDD patterns: early after fire occurrence higher values of the indices were encountered (
Ripley’s
Fire caused a change in the population structure of both forest categories. The two communities were similar in terms of number and size of fire-killed and surviving tree stems. Fire affected mostly the smaller diameters, greatly reducing their number. At the same time, the majority of largest individuals, mainly maritime pines, were killed.
Following disturbance, the density of living stems and snags showed an opposite behaviour as a consequence of self-restoration strategies and fall dynamics. The processes of tree mortality and snag recruitment are balanced by snag decay and fall (
The passage of a high intensity fire front induced a modification in the proportion of sprouters and seeders, due to the concurrent presence of species with varying degrees of resistance and resilience to fire. Regeneration strategies can directly influence species abundance and consequently both dominance and evenness values. Consistently with
Both forest categories showed a generalized increase in structural complexity of living stems immediately following the fire, while in the early post-fire years a decline was registered, reaching values lower that those found for unburned plots. By affecting mostly smaller size stems, fire reduced their pre-disturbance structural dominance. As a consequence, there were more even diameter and height distributions, leading to an increase in structural diversity. This was however a short-lived modification: six years after the perturbation, the high proportional abundance of sprout-origin individuals determined a renewed simplification in structural diversity
Dissimilar patterns of mortality before the disturbance generated a different behavior in the structural diversity of dead stems. In the BC category, characterized by a high presence of chestnut and maritime pine affected respectively by
Fire determined a larger heterogeneity of the vertical structure in the earlier burned plots, enhancing the evenness of cover distribution between the different layers throught the recovery dynamics. The fire event induced an increase in aggregation within the stands. Disturbance by fire has actually been found to increase the degree of clumping (
Segregation mechanisms were generated by the fire, inducing a spatial repulsion between surviving and killed stems. Six years after the fire event, natural restoration dynamics and snag fall processes altered living and dead stems spatial relationships. Close to snags, positive afterlife effects (
Fire frequently creates a high degree of spatial variability in plant survival, which partially depends on the size of individuals (
Fire severity influences the number and the type of plants dying in a fire and their consequent spatial patterns. Fires of low or medium severity produce a highly selective mortality, depending on the species and size of each individual, while fires of high severity usually kill all individuals, regardless of species or size (
Fire severity also influences post-fire regeneration in the burned area (
Seedling recruitment was similar in richness and abundance within both forest categories immediately after fire. In early successional stages after fire, an abundant establishment of saplings and sprouts has been documented by several authors (
Fire effects were found to affect more tree dimensions and spatial patterns rather than composition. Snag dynamics, in particular, were a key element of the observed changes in forest structure, mainly as a consequence of their short-term evolution.
The sprouter-dominated forest stands reveal a major difficulty in restoring the pre-fire conditions. Since the majority of adult conifers in the whole area were killed by the fire, the low number of pine seedlings can cause a partial failure of direct regeneration (
The scarce number of tree seedlings six years after fire occurrence, along with the high resprouting rates and ground cover of shrub species, may suggest a shift towards a community dominated by a low structured, shrub vegetation. A similar scenario has been proposed for the Mediterranean basin with higher disturbance frequency (
Forthcoming changes in fire regimes, yet largely unknown (
Diameter distribution of living stems and snags. Living stems are represented in black, snags in grey. Solid lines correspond to the death ratio,
Second order bivariate Ripley’s
Stand characteristics of the six investigated plots. (BC_nb): not burned broadleaved-coniferous forest stand; (BC_03): broadleaved-coniferous forest stand burned in 2003; (BC_98): broadleaved-coniferous forest stand burned in 1998; (MB_nb): not burned mixed-broadleaved forest stand; (MB_03): mixed-broadleaved forest stand burned in 2003; (MB_98): mixed-broadleaved forest stand burned in 1998.
Plots | BC_nb | BC_03 | BC_98 | MB_nb | MB_03 | MB_98 |
---|---|---|---|---|---|---|
Mean elevation (m a.s.l.) | 277 | 231 | 277 | 236 | 225 | 170 |
Aspect | N-E | N-E | N-E | S-E | S-W | N-W |
Slope (degrees) | 22 | 28 | 22 | 29 | 19 | 17 |
Living stem density (stems ha-1) | 2310 | 182 | 929 | 3787 | 394 | 994 |
Living stem basal area (m2 ha-1) | 15.8 | 5.8 | 2.9 | 16.8 | 7.9 | 4.9 |
Snag density (stems ha-1) | 345 | 2716 | 1421 | 425 | 2643 | 348 |
Snag basal area (m2 ha-1) | 4.3 | 23.8 | 10.1 | 2.1 | 19.9 | 1.1 |
Regeneration density (n ha-1) | 9800 | 11100 | 21900 | 7300 | 11900 | 1600 |
Regeneration richness (n |
5 | 6 | 10 | 7 | 10 | 7 |
Number of top-killed stools and resprouting percentage [Re. (%)] in the four burned plots, for the main resprouting species. (BC_03): broadleaved-coniferous forest stand burned in 2003; (BC_98): broadleaved-coniferous forest stand burned in 1998; (MB_03): mixed-broadleaved forest stand burned in 2003; (MB_98): mixed-broadleaved forest stand burned in 1998.
Species | BC_03 | BC_98 | MB_03 | MB_98 | ||||
---|---|---|---|---|---|---|---|---|
# stools | Re. (%) | # stools | Re. (%) | # stools | Re. (%) | # stools | Re. (%) | |
|
41 | 92.7 | 18 | 44.4 | 31 | 77.5 | 35 | 25.7 |
|
100 | 49 | 72 | 9.7 | 13 | 27.7 | - | - |
|
82 | 72 | 114 | 77.2 | 13 | 34.2 | 12 | 58.3 |
|
74 | 68.9 | 1 | 0 | 87 | 55.4 | 7 | 0 |
|
- | - | 3 | 100 | 33 | 75 | 6 | 0 |
|
- | - | 5 | 60 | 17 | 65.4 | 14 | 7.1 |
Structural diversity in the tree layer through distance-independent variables. Measures were computed for living trees with Brillouin and Vertical Eveness (VE) indices, while tree height diversity (THD) index and tree diameter diversity (TDD) index were calculated for both living stems and snags. (BC_nb): not burned broadleaved-coniferous forest stand; (BC_03): broadleaved-coniferous forest stand burned in 2003; (BC_98): broadleaved-coniferous forest stand burned in 1998; (MB_nb): not burned mixed-broadleaved forest stand; (MB_03): mixed-broadleaved forest stand burned in 2003; (MB_98): mixed-broadleaved forest stand burned in 1998.
Structural diversity | BC_nb | BC_03 | BC_98 | MB_nb | MB_03 | MB_98 |
---|---|---|---|---|---|---|
Richness (n |
10 | 5 | 9 | 7 | 8 | 8 |
Brillouin diversity (HB) | 1.617 | 1.077 | 1.273 | 1.55 | 1.489 | 1.402 |
Brillouin evenness (E) | 0.717 | 0.749 | 0.616 | 0.81 | 0.781 | 0.707 |
THD living stems | 1.469 | 1.838 | 0.795 | 1.28 | 1.887 | 1.026 |
THD snags | 1.516 | 1.487 | 1.122 | 0.966 | 1.716 | 0.668 |
TDD living stems | 1.02 | 1.736 | 0.296 | 0.731 | 1.713 | 0.628 |
TDD snags | 1.314 | 1.063 | 1.009 | 0.8 | 1.119 | 0.367 |
VE | 0.86 | 0.82 | 0.97 | 0.81 | 0.8 | 0.88 |
Ripley’s
Plot | Classes | t(m) | ||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | ||
BC_nb | Living stems | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ |
Snags | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | • | • | • | ◊ | • | |
Total Stems | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | • | ◊ | ◊ | ◊ | |
BC_03 | Living stems | - | - | - | - | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | - | - | - | - | ◊ | - | - | - | - | • | ◊ | ◊ | • | ◊ |
Snags | - | - | • | • | • | • | • | • | • | • | ◊ | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | |
Total Stems | - | - | • | • | • | • | • | • | • | • | ◊ | • | • | • | • | • | ◊ | • | ◊ | ◊ | • | ◊ | • | • | • | |
BC_98 | Living stems | - | • | ◊ | • | • | • | • | ◊ | • | ◊ | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ |
Snags | - | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | |
Total Stems | - | • | • | ◊ | • | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | • | • | ◊ | • | • | • | ◊ | |
MB_nb | Living stems | - | • | • | • | • | • | • | • | ◊ | ◊ | • | ◊ | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | - | - | - |
Snags | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
Total Stems | - | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | - | |
MB_03 | Living stems | • | • | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ |
Snags | - | • | • | • | • | • | • | ◊ | • | ◊ | • | ◊ | ◊ | • | ◊ | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | - | |
Total Stems | - | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | - | - | - | |
MB_98 | Living stems | • | • | • | • | • | • | • | • | • | • | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | • | ◊ | ◊ | • |
Snags | - | - | - | - | - | - | - | - | - | ◊ | ◊ | ◊ | - | - | - | - | - | - | - | - | - | - | - | - | - | |
Total Stems | • | • | • | • | • | • | • | • | • | ◊ | • | ◊ | • | • | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ | ◊ |
Appendix 1 - Diversity indices and the corresponding equations.