Ecosystems and global services : an
outlook on forest and mountain region
Riccardo Valentini
Università della Tuscia
Dipartimento di Scienze dell’Ambiente Forestale
e delle sue Risorse
[email protected]
http://gaia.agraria.unitus.it
Welcome in the Anthropocene !
CO2
CH4
N2O
2007 un anno per il Clima
4° Rapporto Intergovernativo sui Cambiamenti Climatici
Premio Nobel per la Pace
Un film sul clima
Artico si scioglie
Bush torna su i suoi passi ?
Il delfino Baiji è estinto
What was unique?
Ecosystem services
Provisioning
Regulating
Cultural
Goods produced or
provided by
ecosystems
Benefits obtained
from regulation of
ecosystem processes
Non-material
benefits from
ecosystems
Photo credits (left to right, top to bottom): Purdue University, WomenAid.org, LSUP, NASA, unknown, CEH Wallingford, unknown, W. Reid, Staffan Widstrand
Source: NASA
Global C Budget
Atmospheric
accumulation rate
3.2 GtC per year 1990s
Atmosphere
Surface
biosphere
6.3
2.2
F Fuel, Land-Use
Cement Change
Fast process (1 – 102 days)
Gruber et al 2003 , SCOPE project
2.9
Land
Uptake
2.4
Ocean
Uptake
Slow process (103 – 104 days)
4.0
2.0
21
23
12
22
15
-1
-1
NEE (t C ha y )
0.0
18
5
-2.0
7
-4.0
25
17
23
24
6
3
16
13
9
8
2a
4
10
14
11
-6.0
1
20
2
19
-8.0
-10.0
35
40
45
50
55
60
65
70
Latitude (°N)
Valentini, Dolman, Matteucci et al. Nature 2000
VULNERABILITY OF BIOSPHERE
(feed-backs with carbon cycle)
Coupled carbonclimate models
GtCO2/yr
140
CO2-equivalent emissions
120
100
80
60
Baseline A2
BIOSPHERE
40
20
IMAGE S650e
IMAGE S550e
0
1970 1990 2010 2030 2050 2070 2090
Source or sink ?
Vulnerability of Carbon Pools
Carbon in tropical vegetation:
340 Pg
Carbon in wetlands:
450 PgC
Carbon in frozen soils:
400 PgC
• Risk over the coming century of up to 200 ppm of atmospheric CO2
• Not included in most climate simulations.
Gruber et al. 2004
……BIODIVERSITA’ IN CIFRE……
1,7 MILIONI DI SPECIE CONOSCIUTE
15 MILIONI SPECIE STIMATE SULLA TERRA
90% DELLE SPECIE SCONOSCIUTE
Change in Species Diversity
Number per Thousand Species
Number of Species
10000
140
120
1000
100
100
100 to 1000fold increase
North America
Europe
80
60
10
40
1
20
0,1
0
Fossil
Recent
Future
1790-1819 1820-1849 1850-1879 1880-1909 1910-1939 1940-1969 1970-1999
1790
1900
2000
Extinctions
Homogenization
(per thousand years)
(e.g. growth in marine species
introductions)
Source: Millennium Ecosystem Assessment
The experimental site is located in a farm (Malga
Arpaco) at 1699 m a.s.l.
Mean annual temperature: 5 °C
Total annual rainfall: 1200 mm
Soil type: Typic Hapludalfs, fine loamy (FAO)
Ecosystem type: alpine semi-natural grassland
Ecosystem management: extensive
management, pasture from Jun to Sep
Period of EC measurements: 2003-2007
Eddy Covariance type: Metek USA-1, Li-cor 7500
Tower height: 2 m
N2O emission and CH4 uptake was
evaluated fortnightly, during
2003 and 2004 pasture season, using
diffusion chambers. Gas
samples conserved in vacuum vials
were analysed through
gaschromatography
technique.
For the N2O: ECD detector at 320°C;
for the separation a capillary
column Cromosob 1010 at 140°C was
used, with a flux of helium
at 30 kPa. For the CH4: FID detector
at 180°C; for the separation a
column 4m x ¼’’ OD Porapak q 80/100
MESH at 30° was used.
The human foot print
Data
Magnani et al., 2007
Luyssaert et al., submitted
Extreme climate events or disturbances have a strong
effect on biosphere-astmosphere exchanges
Annual mean 1850-2000: 35 M m3 of forest wood
damaged by natural disturbances in Europe.
53% wind throw
16% fire
16% biotic (insects)
3% snow
5% other abiotic
Tatra Experiment CarboEurope
Mean day on monthly base
10
0
EX
IF
NEX
-5
-10
October
September
August
July
June
May
April
Marchr
Februaryr
January
December
-25
November
-20
October
-15
September
Fc [mol m-2 s-1]
5
QUALCHE ESEMPIO
Vannini, Anselmi et al. 2007
Progetto CarboItaly
Malattie epidemiche causate da organismi introdotti
5
0
um
P.
c
ac
t
or
ic
ol
a
ci
tr
P.
ca
m
bi
vo
ra
P.
ci
nn
am
om
i
-5
P.
Phythopthora cinnammomi, uno degli agenti
causali del mal dell’inchiostro del castagno,
è attualmente ristretta a quelle aree in cui la
temperatura minima non scende al di sotto
di 0°C (vedi grafico a destra).
Un aumento delle temperature minime di 24°C, teoricamente verificabile nell’arco di
20-40 anni, porterebbe questa specie ad
espandere il suo areale alle zone castanicole
dove sono oggi presenti specie di
Phytophthora meno aggressive quali P.
cambivora, P. cactorum e P. citricola
La spiccata polifagia di P. cinnamomi,
permetterebbe inoltre al patogeno di
colonizzare nuovi ospiti precedentemente
non raggiungibili per limiti climatici.
Temperature °C
10
QUALCHE ESEMPIO
Vannini, Anselmi et al. 2007
Progetto CarboItaly
Malattie endemiche causate da organismi nativi
35
% d'isolabilamento
Biscogniauxia mediterranea, è un fungo
Ascomycota che vive comunemente come
endofita indifferente all’interno dei tessuti
corticali e legnosi di querce mediterranee.
Durante eventi particolarmente siccitosi,
quando il potenziale idrico fogliare minimo
dell’ospite raggiunge valori inferiori a -2.0
MPa, la popolazione endofitica va
gradatamente aumentando (vedi grafico)
fino a quando, a valori inferiori a -3.0 MPa,
il fungo passa dalla fase endofitica a quella
patogenetica aggredendo rapidamente i
tessuti dell’ospite e causando il cosiddetto
“cancro
carbonioso
delle
querce”.
L’aumento delle temperature estive e la
maggior frequenza di fenomeni estremi, tra
cui la siccità, potrebbero “attivare” un alto
numero di organismi comunemente “silenti”
innescando pericolosi eventi di deperimento
di cenosi forestali
30
25
20
15
10
5
-4,0
-3,5
-3,0
-2,5
MWP (MPa)
-2,0
-1,5
-1,0
Forest patterns
Spatial modelling of forest patterns in dependence by location characteristics
is a reliable way to analyze the possible trajectories and shifts of species
habitat in the near future if environmental conditions will change.
Actual species distribution
Statistical analysis
Driving factors
influencing distribution
Probability of
occurrence
Neighborhood
criteria
Future spatial
distribution
Scenarios of future
driving factors
Actual species distribution
Calibration
Statistical
analysis
Driving factors
influencing distribution
Probability
of occurrence
Future
Spatial Distribution
Physiognomic categories
Scenarios future
driving factors
Forest Map of Italy (1:100000)
raster 250 meters of resolution
Error in
rasterization
-0.15%
%
00 - Woody plantation in agricultural areas
0.55
01 - Oaks and other evergreen broadleaf forests
9.07
02 - Deciduous oak-dominant forests
24.45
03 - Chestnut-dominant forests
8.85
04 - Beech-dominant forests
11.52
05 - Hygrophyte species-dominant forests
0.85
06 - Other broadleaf deciduous autochthon species-dominant forests
10.28
07 - Exotic broadleaf-dominant forests and plantations
1.85
08 - Mediterranean pine and cypress dominant forests
2.46
09 - Oro-Mediterranean and mountain pine dominant forests
2.75
10 - Abies alba and Picea rubens dominant forests
7.71
11 - Larch and cembrus pine dominant forests
3.06
12 - Exotic needleleaf dominant forests
0.10
13 - Mixed needleleaf and broadleaf forests with prevalent beech
2.19
14 - Mixed needleleaf and broadleaf forests with prevalent oro-mediterranean and mountain pine
2.24
15 - Mixed needleleaf and broadleaf forests with prevalent Abies alba and/or Picea rubens
1.93
16 - Mixed needleleaf and broadleaf forests with other species prevalent
6.77
17 - Tall Mediterranean Macchia
3.35
26% of Italian territory is forest
Actual species distribution
Calibration
Statistical
analysis
Driving factors
influencing distribution
Probability
of occurrence
Future
Spatial Distribution
Scenarios future
driving factors
Driving factors
•Elevation values (m above sea level)
•Slope value (°)
•Aspect value (° clockwise from north)
•Mean annual precipitation (mm)
•Mean annual snow water equivalent (mm)
•Mean daily short wave net radiation (W/m2)
•Mean of the annual dew point temperature (°K)
•Mean of the minimum annual temperature (°K)
•Mean of the maximum annual temperature (°K)
DEM srtm
DMI F12 A2
Actual species distribution
Calibration
Statistical
analysis
Driving factors
influencing distribution
Logistic regression
Probability
of occurrence
Future
Spatial Distribution
Scenarios future
driving factors
 Pi
log 
 1  Pi

   0  1  x1   2  x2   3  x3  ...   n  xn

where Pi is the probability for the occurrence of the considered forest type on location i and the x's are the location factors
(independent variable values) forcing the presence/absence of forest classes.
ROC Curve
Accuracy
1.0
Sensitivity
0.8
0.6
0.4
ROC 0.973
0.2
0.0
0.0
0.2
0.4
0.6
0.8
1.0
1 - Specificity
i.e.ROC curve test for class 8
Diagonal segments are produced by ties.
Mean ROC 0.855
Actual species distribution
Calibration
Statistical
analysis
Driving factors
influencing distribution
Probability
of occurrence
Neighbooring
criteraia
Future
Spatial Distribution
Scenarios future
driving factors
Example of Euclidean distance grid
Example of distance-based probability grid
Piv
Pi  wis  Pi s  wi v Piv
Forest classes
00 - Woody plantation in agricultural areas
01 - Oaks and other evergreen broadleaf forests
09 - Oro-Mediterranean and mountain pine dominant forests
Altitude profiles of forest distribution
10 - Abies alba and Picea dominant forests
class 0
3000
class 1
class 2
2500
11 - Larch and cembrus pine dominant forests
class 3
class 4
class 5
2000
1500
m asl
class 6
12 - Exotic needleleaf dominant forests
class 7
class 8
class 9
13 - Mixed needleleaf and broadleaf forests with prevalent beech
class 10
1000
class 11
class 12
500
14 - Mixed needleleaf and broadleaf forests with prevalent oro-mediterranean and mountain pine
class 13
class 14
class 15
0
0
500
1000
1500
2000
2500
class 16
3000
Number of pixels
Actual distribution
3000
2500
class 2
class 3
class 4
1500
m asl
2000
class 5
class 6
class 7
class 8
1000
class 9
class 14
500
Case a) Changed areas (red, 82%) considering only
statistical analysis
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
class 15
class 16
0
5500
Number of pixels
Case a)
3000
class 0
2500
class 2
class 3
class 4
2000
m asl
class 5
1500
class 6
class 7
class 8
1000
class 9
class 10
500
class 14
class 15
class 16
0
500
1000
1500
2000
2500
3000
Number of pixels
Case b) Changed areas (red, 77%) considering
statistical analysis and neighborhood criteria
3500
4000
4500
5000
0
5500
CONCLUSIONS
• Climate change will impact mountain
ecosystems in different and possible
unexpected ways (increase productivity,
decrease biodiversity…)
• The human dimension is still important
• Conservation of old forests preserve
ecosystem services
“You can observe a lot, just by watching.”
-Yogi Berra
Thank You