The Vegetation Resilience After Fire (VRAF) index: Development, implementation and an illustration from central Italy

https://doi.org/10.1016/j.jag.2007.12.003Get rights and content

Abstract

A suitable index is proposed to evaluate the natural short–medium-term recovery capability of vegetation in burnt areas. The study area covers 2450 km2 in western Tuscany (Province of Pisa, Italy). This region is characterized by a typical Mediterranean climate and is subject to fire damage during the dry summer season. Damage is mitigated where a natural rapid regrowth of vegetation prevents soil erosion, supporting the return to a natural pre-fire state.

The Vegetation Resilience After Fire (VRAF) index is based on the vegetation association, soil type and geology, and on morphological features such as slope and aspect. The results are proposed as georeferenced maps defining areas with different vegetation resilience for both high and medium burn severity. The VRAF maps estimate the natural ability of vegetation to recover after fire, and suggest where human intervention is required to improve this capability. The VRAF index was checked by monitoring vegetation regrowth after fire in three burnt areas over a five-year period using spectral signatures, the feature space and the NDVI derived from remote sensing data. This analysis indicates that the high values of the VRAF index correspond to a recovery period of almost three years. Field surveys were performed to further test the results. On the whole, the VRAF index is a good parameter for assessing the capability of vegetation to recover in northern Mediterranean areas.

Introduction

In the Mediterranean region, fire is the major cause of ecosystem damage during the summer, and countries use many resources to extinguish fires and recover burnt areas. Nevertheless, wildfires are a natural occurrence in Mediterranean environments, therefore it show a high post-fire regeneration capacity (Naveh, 1974, Trabaud et al., 1985, Thanos et al., 1989, Thanos and Marcou, 1991).

As woodlands cover 35% of the Province of Pisa (western Tuscany), fires constitute an important hazard; in fact more than 453 ha of forested area was burnt in the 2000–2004 period, for a total of 246 forest fires (AIB, 2001–2004).

GIS applications are largely used in natural hazard management (Le Bissonnais et al., 2001, Jaiswal et al., 2002, Yoshino and Ishioka, 2005), and although previous studies have addressed the problem of fire in Tuscany (e.g. Maselli et al., 2003), they have mainly focused on fire risk.

Fire consumes the protective vegetation and organic litter cover, which can destabilize surface soils on steep slopes (Shakesby and Doerr, 2006). Immediately after a fire and during the following months, the infiltration significantly decreases while surface erosion increases because the bare soil is exposed to raindrop impact and surface runoff (Inbar et al., 1998, Giovannini et al., 2001). These impacts vary with the severity of the fire, soil property and post-fire climatic condition (Certini, 2005). The natural vegetation regeneration after fire stabilizes the burnt area (Beyers, 2004), therefore recovery of vegetation seems to be the main factor limiting the fire damages and its consequences. (Inbar et al., 1998, De Luis et al., 2001, Cerdà and Doerr, 2005). The months immediately after a fire are the most critical for the reconstruction of adequate soil and plant mantles (Pardini et al., 2004); therefore, plant regeneration just after fire, which reduces the risk of erosion by rains, is one of the most important elements assisting the return to pre-fire soil–plant conditions. Pausas et al. (1999) report about one and three years for recovering 50 and 60% of burnt area, respectively, Calvo et al. (2003) stated that the percentage of bare soil decreases very rapidly first year after the fire, principally for the herbaceous growth, and three years after the fire for the woody species, Andreu et al. (2001) shows that the soil losses are higher during the four months after the fire and Cerdà et al. (1995) stated a very strong reduction in the mean erosion rate two years after fire. Considering these works, five years seems to be a reasonable time to evaluate the response of soil–plant system after the fire.

We present the Vegetation Resilience After Fire (VRAF) index to evaluate the short–medium term (five years) recovery capability of vegetation (vegetation resilience) after fires of high and medium burn severity, respectively. The index is computed with a factor-based equation combining five different parameters and their relative weights, assuming that the dominant community species remain the same after fire (Broncano et al., 2005). The index was tested by remote sensing analysis of three sites affected by fire in the 2001–2003 period and by field observation of four sites burnt in summers 2001, 2003 and 2004. Burn severity was estimated using the delta Normalized Burn Ratio (dNBR) in the remote sensing procedure (Key and Benson, 1999, Key and Benson, 2006, Miller and Yoll, 2002, van Wagtendonk et al., 2004, Epting et al., 2005, Lentile et al., 2006), while regeneration after fire was evaluated through multitemporal spectral signatures and the Normalized Difference Vegetation Index (NDVI) (Kasischke and French, 1995, Viedma et al., 1997, Riano et al., 2002, Diaz-Delgado et al., 2003, Maselli et al., 2003), as well as the feature space (Galvao and Vitorello, 1998).

The final maps, obtained by applying VRAF to both high and medium burn severity cases, provide useful indications for planning restoration actions and post-fire resource management.

Section snippets

The study area

The study area covers the entire Province of Pisa (∼2450 km2) localized in western Tuscany (Fig. 1). The orography landscape is characterized by low gentle slopes except southeast of Mt. Pisano, where the maximum elevation is 917 m a.s.l.

The study area has a typical Mediterranean climate with 2001–2004 average precipitation of ∼200 mm/yr−1 mainly concentrated during the autumn and spring season [http://www.scia.sinanet.apat.it]. Periods of highest temperatures and lowest precipitation coincide

Geographic database

In order to map the short–medium term recovery capacity of vegetation after fire in the woodlands of a typical Mediterranean coastal region, we propose a suitable index named VRAF based on five interacting parameters: soil type, vegetation cover, slope, aspect and geology. These parameters are organized in a geographic database georeferenced (WGS84, UTM Zone 32 coordinate system), summarized in Fig. 2, where geographic layers are stored both in raster (cell size 10 m) and vector format. We did

Testing the VRAF index

The VRAF index was validated by using remote sensing data and field surveys in four sites affected by fire during the 2001–2004 period (Fig. 1). Four sites were not enough to rigorously validate the VRAF index, however they can give useful indications on its reliability. The sites are named PSA1, PSA2, PSA3, and PSA4 (Table 3). Since the maximum distance between sites is less than 25 km, they can be considered to receive similar rainfall.

Zones PSA1, PSA2 and PSA3 were affected by the oldest

Results and discussion

For all the vegetated areas of the Province of Pisa the VRAF maps show: (i) a generally high ability to recover from fire in the alluvial plain near the coast and in the Valdarno and Valdera areas (Fig. 4, Fig. 5). In these localities the dominant associations are Quercus and Pinus, deciduous oak coppices and pine forests, the soil mainly consists of coastal sandy soils, loose alluvial soils, old alluvial soils and hilly sandy soils, and the slope is gentle; (ii) a moderate to low ability to

Conclusions

Fire is the most significant natural hazard for forests areas of the Mediterranean region. Fire management entails the rehabilitation of a burnt area in order to mitigate the negative effects of fire. Mediterranean areas are characterized by heavy autumn rains; at this time of year, if the forest canopy on steep slopes has been damaged or completely destroyed by fire, there is an elevated risk of erosion which may be mitigated by rapid vegetation regrowth soon after the fire.

In this context,

References (75)

  • K.J. Elliott et al.

    Vegetation dynamics after a prescribed fire in the southern Appalachians

    Forest Ecol. Manage.

    (1999)
  • J. Epting et al.

    Evaluation of remotely sensed indices for assessing burn severity in interior Alaska using Landsat TM and ETM+

    Remote Sens. Environ.

    (2005)
  • S. Fernandez et al.

    A susceptibility model for post-wildfire soil erosion in a temperate oceanic mountain area of Spain

    Catena

    (2005)
  • L.S. Galvao et al.

    Variability of laboratory measured soil lines of soil from southeastern Brazil

    Remote Sens. Environ.

    (1998)
  • M.A. Gilabert et al.

    A generalized soil-adjusted vegetation index

    Remote Sens. Environ.

    (2002)
  • G. Giovannini et al.

    Effect of land use and eventual fire on soil erodibility in dry Mediterranean conditions

    Forest Ecol. Manage.

    (2001)
  • M. Inbar et al.

    Runoff and erosion processes after a forest fire in Mount Carmel, a Mediterranean area

    Geomorphology

    (1998)
  • R.K. Jaiswal et al.

    Forest fire risk zone mapping from satellite imagery and GIS

    Int. J. Appl. Earth Observ. Geoinf.

    (2002)
  • K.H. Kim et al.

    Effect of rainfall electrolyte concentration and slope on infiltration and erosion

    Soil Technol.

    (1996)
  • C. Kosmas et al.

    The effect on rock fragments on wheat biomass production under highly variable moisture condition in Mediterranean environments

    Catena

    (1994)
  • L. Kumar et al.

    Relationship between variation growth rates at the onset of wet season and soil type in the Sahel of Burkina Faso: implication for resource utilization at large scale

    Ecol. Model.

    (2002)
  • J. Mataix-Solera et al.

    Hydrophobicity and aggregate stability in calcareous topsoils from fire affected pine forests in southern Spain

    Geoderma

    (2004)
  • F. Maselli et al.

    Use of NOAA-AVHRR NDVI images for the estimation of dynamic fire risk in the Mediterranean areas

    Remote Sens. Environ.

    (2003)
  • J.D. Miller et al.

    Mapping forest post-fire consumption in several overstay types using multi-temporal Landsat TM and ETM data

    Remote Sens. Environ.

    (2002)
  • A.A. Millward et al.

    Adapting the RUSLE to model soil erosion potential in a mountainous tropical watershed

    Catena

    (1999)
  • Z. Naveh

    Effects of fire in Mediterranean Region

  • G. Pardini et al.

    Relative influence of wildfire on soil properties and erosion processes in different Mediterranean environments in NE Spain

    Sci. Total Environ.

    (2004)
  • D. Riano et al.

    Assessment of vegetation regeneration after fire through multitemporal analysis of AVIRIS images in the Santa Monica Mountains

    Remote Sens. Environ.

    (2002)
  • T.L. Saaty

    A scaling method for priorities in hierarchical structures

    J. Math. Psychol.

    (1977)
  • R.A. Shakesby et al.

    Wildfire as a hydrological and geomorphological agent

    Earth Sci. Rev.

    (2006)
  • A.D. Thomas et al.

    Nutrient losses in eroded sediment after fire in eucalyptus and pine forests in the Mediterranean environment of Northern Portugal

    Catena

    (1999)
  • G.M.E. van der Merwe et al.

    Factors that govern the formation of melanic soils in South Africa

    Geoderma

    (2002)
  • J.W. van Wagtendonk et al.

    Comparison of AVIRIS and Landsat ETM+ detection capabilities for burn severity

    Remote Sens. Environ.

    (2004)
  • O. Viedma et al.

    Modelling rates ecosystem recovery after fire by using Landsat TM data

    Remote Sens. Environ.

    (1997)
  • J.H. Witty et al.

    Contribution of water supply from the weathered bedrock zone to forest soil quality

    Geoderma

    (2003)
  • S.M. Wondzell et al.

    Postfire erosional processes in the Pacific Northwest and Rocky Mountain regions

    Forest Ecol. Manage.

    (2003)
  • AIB (Anti Incendio Boschivo) schede, 2001–2004. Regione...
  • Cited by (41)

    • Vegetation cover change and restoration potential in the Ziwuling Forest Region, China

      2023, Ecological Engineering
      Citation Excerpt :

      The effect of slope on the recovery potential of vegetation was small, and climatic factors had a greater influence, which is consistent with Lv et al.'s finding that slope in central and northern Mongolia has a small impact on vegetation restoration potential (Lv et al., 2021). However, Bisson et al. (2008) found that it was inversely proportional to the damage suffered, and slope had a significant effect on the vegetation recovery capacity – areas with lower slopes are more resilient. This may be because the surface vegetation in the Ziwuling Forest Region is in good condition, and it can effectively retain water in the steep slope area.

    • Development of post-fire vegetation response-ability model in grassland mountainous ecosystem using GIS and remote sensing

      2020, ISPRS Journal of Photogrammetry and Remote Sensing
      Citation Excerpt :

      This exposure will lead to an increase in surface erosion. Post-fire vegetation regeneration is a fundamental process because it mitigates the negative impacts of fire (Bisson et al., 2008). In other words, vegetation regeneration supports the return to pre-fire conditions and decreases erosion, alteration of plant species composition by invasive species, and vegetation structure.

    • Evaluating the vegetation restoration potential achievement of ecological projects: A case study of Yan'an, China

      2020, Land Use Policy
      Citation Excerpt :

      We define this kind of potential as the similar habitat based vegetation restoration potential (SHBVSP). Differences in SHBVSP lead to differences in the difficulties of vegetation restoration; the greater the SHBVSP, the easier it is to restore vegetation, and vice versa (Bisson et al., 2008; Arianoutsou et al., 2011). To evaluate vegetation restoration in the gully region of the Loess Plateau, Gao et al. (2017) used soil, topography, and meteorological condition indicators to delineate similar habitat units, measured their vegetation coverage using SPOTVEG NDVI data, and then used statistical methods and geospatial analysis techniques to build the SHBVSP model.

    View all citing articles on Scopus
    View full text