Final Report Summary - RESCRI (Shifts in ecosystems state in Mediterranean landscapes: when, where and how? The interacting effects of multiple disturbances under climate change)
The objective of the RESCRI project was to increase our capability in predicting changes in ecosystem state over Mediterranean ecosystems (MTEs) as a result of altered regimes in multiple disturbances under altered climates. Concretely, the project aimed at describing how vegetation characteristics (e.g. regenerative strategy, succession rate, flammability) and the interacting effects of two major disturbances in MTEs (fire and drought) govern the dynamics and properties of Mediterranean landscapes. The major objectives of the project were (a) the development of a novel model explicitly including the interacting effects of fire and drought, and (b) the evaluation of resilience and critical behavior (sudden shifts in ecosystem state) through extensive scenario-based simulations. More specifically, the project: (i) examined the influence of interacting disturbances on vegetation composition relative to individual disturbance regimes; (ii) tested whether the effects of interacting fire and drought on vegetation composition are synergistic or antagonistic; (iii) evaluated the influence of simultaneous and consecutive disturbance events as mechanisms contributing to observed vegetation dynamics; (iv) assessed the relative dynamics and abundance of major regenerative strategies in MTEs –seeders and resprouters–; (v) determined the likelihood of sudden vegetation changes from forest to non-forest dominated landscapes; and (vi) identified specific sequences of events that have the capacity to promote non-linear dynamics in Mediterranean ecosystems.
The developed model was implemented in R statistical language (R Core Team 2015) and corresponds to a state and transition simulation model (STSM) integrated in a landscape of 400 (20 × 20) cells where vegetation dynamics is influenced by stochastic regimes of fire (implemented as a spatially explicit process) and drought. Each cell is described by the following state variables: the proportion of six major vegetation types, their corresponding live and dead biomass values, and a ‘counter’ of time since disturbance that modulates processes of cells’ state change. Vegetation types are defined by broad formations, constituted by sparse vegetation, shrublands, and two developmental stages of a tree seeder and a tree resprouter. According to major MTEs ecological characteristics, the different vegetation types show different sensitivity to fire and drought. The spatial scale of the model is not explicitly fixed, but it would correspond to a regional landscape affected by major synoptic weather conditions (e.g. drought episodes) where each cell would represent an area large enough (e.g. 0.5–1 square km) to encompass a mosaic of vegetation types with shared environmental conditions and capabilities to respond to the ecological processes incorporated in the model. Vegetation processes are implemented cell by cell, but each cell’s dynamics is coupled with the rest of cells both by disturbance regimes, and by the influence of landscape properties (e.g. fraction of mature forests, biomass) on the rate of vegetation succession and replacement processes or the likelihood of fire.
Scenario-based modelling experiments were performed to evaluate how changes in fire and drought regimes (and their interaction) may undermine the resilience of MTEs. Two distinct scenario-based approaches were implemented.
- First, three contrasting regimes for fire and three regimes of increasing drought recurrence were defined. Each of these six regimes was applied in a corresponding individual disturbance scenario, and nine additional compound scenarios resulted from all the possible combinations of fire and drought regimes. A no disturbance scenario was implemented as control. This approach allowed testing the effects of interacting fire and drought –synergism vs. antagonism– following the additive model, in which the effects of interacting disturbances are compared with the additive sum of effects of the corresponding disturbance regimes acting in isolation. Fire scenarios corresponded to: (i) large infrequent fires at a mean frequency of 100 years, (ii) small frequent fires at a mean frequency of 20 years, and (iii) a mixed fire regime including the combination of large infrequent fires and small frequent fires. Drought scenarios corresponded to drought return intervals of 40, 30, and 15 years.
Results from these simulations indicate that the interaction between moderate fire and drought recurrence can act as a strong mechanism generating highly heterogeneous, mosaic-like landscapes in which different regeneration types coexist, as observed in MTEs. Resprouters dominated under individual, moderate disturbance regimes of fire or drought, whereas the interaction of the two disturbances promoted the long-term coexistence of both tree regeneration strategies. However, shrubland expansion and persistence at the expanse of forests was favoured by increases in drought recurrence and associated fire-drought interactions, suggesting important vegetation changes in MTEs under climate change. Overall, the cumulative effects of fire and drought can lead to distinct landscape configurations under moderate disturbance regimes that are otherwise only attained under high frequency of individual disturbances. At the ecosystem level, however, results suggest that disturbance-induced vegetation dynamics can act as a buffer against the interacting effects of fire and drought, precluding the prevalence of synergistic effects of the two disturbances.
- Second, modelling simulations across an extensive set of fire and drought recurrences were used to assess the likelihood of sudden shifts in vegetation composition (critical ecosystem behaviour). Fire and drought recurrences spanning from 15 to 200 years were tested in individual and all the possible compound fire-drought regimes (N = 480 scenarios). The implemented fire regimes corresponded to mixed regimes of small frequent and large infrequent fires. These simulations revealed that compound fire-drought disturbance regimes that sustain mixed forest-shrubland landscapes can suddenly shift to systems dominated by shrublands. Such shifts may be triggered by specific sequences of events with a short interval between disturbances. Large fire events followed by drought years promoted the vast majority of sudden shifts from forest to shrubland; drought conditions after large crown fires would hamper the successful regeneration of tree resprouters and seeders, as juveniles are more sensitive to drought than mature individuals. Subsequent fire-vegetation feedbacks (e.g. increased landscape flammability as a result of shrublands dominance) could preclude the succession to forested states until a window of opportunity (e.g. long inter-disturbance period) favours transition from shrubland to forests. On the other hand, results strongly suggest that the interacting effects of fire and drought regimes effectively modulate the resilience of the system and the potential tipping points, especially in relation to the characteristics of disturbance regimes (e.g. disturbance frequency). For instance, low drought recurrences increase the likelihood of sudden ecosystems shifts under frequent fire regimes, whereas high drought recurrences can promote sudden changes in ecosystem state under regimes of infrequent fires.
The RESCRI project represents a significant advance in disturbance ecology, especially in relation to the effects of disturbance interactions to vegetation dynamics and ecosystem resilience. The simulation experiments performed aid in understanding the basic ecological mechanisms constraining system behaviour and the occurrence of alternate states in Mediterranean ecosystems. As such, results of the RESCRI project could be relevant for the optimization of conservation strategies and maintenance of ecosystem services in Mediterranean areas. Additionally, given that the effects of changing climatic extremes and altered disturbance regimes will affect all types of ecosystems, the framework developed here can be useful to increase the predictability of the effects of global change across a wide range of ecosystems.
The developed model was implemented in R statistical language (R Core Team 2015) and corresponds to a state and transition simulation model (STSM) integrated in a landscape of 400 (20 × 20) cells where vegetation dynamics is influenced by stochastic regimes of fire (implemented as a spatially explicit process) and drought. Each cell is described by the following state variables: the proportion of six major vegetation types, their corresponding live and dead biomass values, and a ‘counter’ of time since disturbance that modulates processes of cells’ state change. Vegetation types are defined by broad formations, constituted by sparse vegetation, shrublands, and two developmental stages of a tree seeder and a tree resprouter. According to major MTEs ecological characteristics, the different vegetation types show different sensitivity to fire and drought. The spatial scale of the model is not explicitly fixed, but it would correspond to a regional landscape affected by major synoptic weather conditions (e.g. drought episodes) where each cell would represent an area large enough (e.g. 0.5–1 square km) to encompass a mosaic of vegetation types with shared environmental conditions and capabilities to respond to the ecological processes incorporated in the model. Vegetation processes are implemented cell by cell, but each cell’s dynamics is coupled with the rest of cells both by disturbance regimes, and by the influence of landscape properties (e.g. fraction of mature forests, biomass) on the rate of vegetation succession and replacement processes or the likelihood of fire.
Scenario-based modelling experiments were performed to evaluate how changes in fire and drought regimes (and their interaction) may undermine the resilience of MTEs. Two distinct scenario-based approaches were implemented.
- First, three contrasting regimes for fire and three regimes of increasing drought recurrence were defined. Each of these six regimes was applied in a corresponding individual disturbance scenario, and nine additional compound scenarios resulted from all the possible combinations of fire and drought regimes. A no disturbance scenario was implemented as control. This approach allowed testing the effects of interacting fire and drought –synergism vs. antagonism– following the additive model, in which the effects of interacting disturbances are compared with the additive sum of effects of the corresponding disturbance regimes acting in isolation. Fire scenarios corresponded to: (i) large infrequent fires at a mean frequency of 100 years, (ii) small frequent fires at a mean frequency of 20 years, and (iii) a mixed fire regime including the combination of large infrequent fires and small frequent fires. Drought scenarios corresponded to drought return intervals of 40, 30, and 15 years.
Results from these simulations indicate that the interaction between moderate fire and drought recurrence can act as a strong mechanism generating highly heterogeneous, mosaic-like landscapes in which different regeneration types coexist, as observed in MTEs. Resprouters dominated under individual, moderate disturbance regimes of fire or drought, whereas the interaction of the two disturbances promoted the long-term coexistence of both tree regeneration strategies. However, shrubland expansion and persistence at the expanse of forests was favoured by increases in drought recurrence and associated fire-drought interactions, suggesting important vegetation changes in MTEs under climate change. Overall, the cumulative effects of fire and drought can lead to distinct landscape configurations under moderate disturbance regimes that are otherwise only attained under high frequency of individual disturbances. At the ecosystem level, however, results suggest that disturbance-induced vegetation dynamics can act as a buffer against the interacting effects of fire and drought, precluding the prevalence of synergistic effects of the two disturbances.
- Second, modelling simulations across an extensive set of fire and drought recurrences were used to assess the likelihood of sudden shifts in vegetation composition (critical ecosystem behaviour). Fire and drought recurrences spanning from 15 to 200 years were tested in individual and all the possible compound fire-drought regimes (N = 480 scenarios). The implemented fire regimes corresponded to mixed regimes of small frequent and large infrequent fires. These simulations revealed that compound fire-drought disturbance regimes that sustain mixed forest-shrubland landscapes can suddenly shift to systems dominated by shrublands. Such shifts may be triggered by specific sequences of events with a short interval between disturbances. Large fire events followed by drought years promoted the vast majority of sudden shifts from forest to shrubland; drought conditions after large crown fires would hamper the successful regeneration of tree resprouters and seeders, as juveniles are more sensitive to drought than mature individuals. Subsequent fire-vegetation feedbacks (e.g. increased landscape flammability as a result of shrublands dominance) could preclude the succession to forested states until a window of opportunity (e.g. long inter-disturbance period) favours transition from shrubland to forests. On the other hand, results strongly suggest that the interacting effects of fire and drought regimes effectively modulate the resilience of the system and the potential tipping points, especially in relation to the characteristics of disturbance regimes (e.g. disturbance frequency). For instance, low drought recurrences increase the likelihood of sudden ecosystems shifts under frequent fire regimes, whereas high drought recurrences can promote sudden changes in ecosystem state under regimes of infrequent fires.
The RESCRI project represents a significant advance in disturbance ecology, especially in relation to the effects of disturbance interactions to vegetation dynamics and ecosystem resilience. The simulation experiments performed aid in understanding the basic ecological mechanisms constraining system behaviour and the occurrence of alternate states in Mediterranean ecosystems. As such, results of the RESCRI project could be relevant for the optimization of conservation strategies and maintenance of ecosystem services in Mediterranean areas. Additionally, given that the effects of changing climatic extremes and altered disturbance regimes will affect all types of ecosystems, the framework developed here can be useful to increase the predictability of the effects of global change across a wide range of ecosystems.