Potential impact of climate change on length of ign ition danger season in Mediterranean shrubland of North SardiniaGrazia Pellizzaro1, Bachisio Arca1, Andrea Ventura1, Pierpaolo Duce1, Valentina Bacciu2, Enrico Scoccimarro2

1 National Council of Research, Institute of Biometeorology (CNR-IBIMET), Sassari, Italy2CMCC Euro-Mediterranean Center for Climate Changes, Italy

The main aim of this work is to identify useful too ls to forecast impacts of expected climate change on live fuel moisture content (Live FMC) in Mediter ranean shrublands.

Moisture content of fine live fuel (diameter < 0.6 cm) was determined periodically during four years on several Mediterranean shrub dominant species of the study area : Cistus monspeliensis L., Pistacialentiscus L., and Juniperusphoenicaea L.Meteorological variables were also recorded.

The potential climate change impact on ignition dan ger season in Mediterranean Basin shrubland was simulated using the Drought code and future climate scenarios at local scale derived from an advanced high resolution Regional Climate Model realised by C MCC and University of Belgrade (SITRA project).

In order to compare different years DC was expresse d as percentage of his maximum value during the differen t seasons. The thresholds for estimating the start an d the end of ignition danger season were calculated by an alysingtwo different data series: data from beginning of y ear until the date in which the drought code reached his maxi mum value were analysed for estimating the start date, w hereas, the others were used for estimating the end of dang er season.

An increase of length of danger fire season was obs erved for scenario A1B.The first results reported here suggest that future climate projections could determine changes in fuel moistur e dynamics and affect the duration of the danger fire season. As a consequence of this fact a probable general in crease of fire danger in the Mediterranean area could be e xpected.

Relation between Drought code and live fuel moistur e content

Expected impacts of climate change on fuel statusDifferences in number of days between Starting and Ending dates estimated using data from the Regional Climate Model for baseline(1961 – 1990) and A1B scenario (2071 – 2100)

Live fuel moisture content annual pattern

The 90th percentile value is about 60% of maximum valuereached during the different seasons for ending date and about 40 % for starting date

The study was carried out in North Western Sardinia, Italy (40°36’ N; 8°09’ E, 30 m a.s.l.). The study area is covered by Mediterranean maquis and garigue. The climate is Mediterranean.

-5

0

5

10

15

20

Day

s

-3

+15

Starting date

Ending date

0

50

100

150

200

250

81 101

117

129

151

165

179

193

207

221

242

256

270

284

298

319

339

Day of the Year

Live

FM

C (%

)R

ainf

all (

mm

)

01002003004005006007008009001000

Dro

ught

cod

e

rainfall Cistus Rosmarinus Drought code

2005

0

50

100

150

200

250

8 49 77 105

136

161

190

239

266

294

322

Day of the Year

Live

FM

C (%

)R

ainf

all (

mm

)

01002003004005006007008009001000

Dro

ught

cod

e

Rainfall Cistus Rosmarinus Drought code

2008

Dependent variable

AIC Akaike Information

Criterion

P-value Hosmer Lemeshow Chi-

Square

% Correct classification

Cistus 62.82 0.65** 82.8

Juniperus 76.93 0.22** 81.7

Pistacia 44.82 0.93** 90.3

Rosmarinus 54.19 0.25** 84.9

Different ranges and seasonal trends of live fuel moisture were observed for the species examined. The highest variability was observed for Cistus and Rosmarinus.

The logistic regression approach was used in order to establish the relationship between the danger season and DC. The danger season was classified as a binary response (true/false) by using a threshold of LFMC of 100%.

The DC explains a large amount of the variability i n the variable drought season. The model based on DC provided sign ificant values of the Hosmer and Lemeshow test, indicating no evidence of a lack of fit; in addition, the models provided highvalues of correct classification for all the specie s. Estimating starting and ending dates of danger seas on

0

50

100

150

200

250

1 15 29 43 57 71 85 99 113

127

141

155

169

183

197

211

225

239

253

267

281

295

309

323

337

351

365

Day of the Year

Live

FM

C (

%)

Cistus 05 Cistus 07 Cistus 08 Cistus 09Rosmarinus 05 Rosmarinus 07 Rosmarinus 08 Rosmarinus 09

0

50

100

150

200

250

1 15 29 43 57 71 85 99 113

127

141

155

169

183

197

211

225

239

253

267

281

295

309

323

337

351

365

Day of the Year

Live

FM

C (

%)

Juniperus 05 Juniperus 07 Juniperus 08 Juniperus 09 Pistacia 05Pistacia 07 Pistacia 08 Pistacia 09

Statistical parameters provided by the logistic ana lysis

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80 90 100% DC respect of maximum value

Per

cent

ile

Starting date

Percentile curve for Drought code considering live fuel moisture data higherthan 100%.

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80 90 100% DC respect of maximum value

Per

cent

ile

LFMC > 100%All species

LFMC > 100%All species

Ending date