Vegetation Dynamics in JULES: The TRIFFID model.

Vegetation Dynamics in
JULES v2.0:
The TRIFFID model.
Rosie Fisher
Stephen Sitch
David Galbraith
Chris Huntingford

TRIFFID history
TRIFFID is the global vegetation model
embedded in the Hadley Centre GCM & JULES.
TRIFFID is similar (not identical) to the majority
of DGVMs in terms of physiology & PFT
composition.
Conceptually different in terms of vegetation
competition & half-hourly hourly gas exchange.

TRIFFID Structure in JULES
Soils
Growth
Phenology (daily)
MOSES
NPP
PFT LAI +
Coverage
TRIFFID
Competition
Climate
Disturbance
Decomposition
Turnover
30-180 minutes 10-30 days

TRIFFID and PFT competition
NL
tree
BL
tree
C3
C4
Change in vegetation carbon
Area increase NPP Litter
Bare Gd
Shrub
Expansion of PFT area: Lotka Volterra

PFT competition
Competition parameters C ij
Define how PFT ‘i’ affects PFT ‘j’
‘Tree – shrub – grass’ dominance
hierarchy.
These parameters are difficult to define…

Groups working with TRIFFID
Hadley Centre JCHMR (HADGEM3)
• First implementation of C-cycle C
in standard
Hadley Model
Reading – soil moisture parameters
CLASSIC – snow modelling, physiology
DGVM intercomparisons – Sitch,
Friedlingstein, Cramer.

Results from TRIFFID
Inter-comparison with other DGVMs in
IMOGEN: Stephen Sitch
Disaggregation of water and temperature
responses in Amazonia: David Galbraith
Acclimation of respiration to increasing
temperature: Owen Atkin & Rosie Fisher

.1. DGVM inter comparison
Sitch et al 2007
Compared 5 DGVMs with the same climate
drivers within the IMOGEN framework.
Previous inter-comparisons (C 4 MIP) have used
different GCM-DGVM pairs. This is a direct
comparison
Compare predicted change in land carbon over
21 st century.
Separate out the CO 2 and climate response.

Sitch et al. 2007

Conclusions:
TRIFFID is much more sensitive to
climate changes than the ‘mean’
model response.
Why is this?

2. Interactions between climate drivers using
MOSES-TRIFFID
(40% Precip Reduction to 2100)
Slides by David Galbraith
160
Change in Amazonian
Vegetation Carbon (PgC)
140
120
100
80
60
40
20
2020 2040 2060 2080 2100
Year
Precipitation
Precipitation + CO 2
Precipitation + Temperature
Precipitation + VPD
Precipitation + CO2 + VPD + Temperature

Interactions with other climate drivers:
Partitioning the ‘dieback’ response
60
40
Which aspect of
temperature is
causing the dieback?
Variance Explained (%)
20
0
-20
-40
-60
Respiration increase
or photosynthetic
decline?
-80
-100
MOSES-TRIFFID HYLAND LPJ
MODEL
VPD
Soil moisture
CO2
TEMP

3. Acclimation of respiration to temperature.
Atkin O, Zaragoza-Castells
J, Fisher R et al. in review
100
50
a
c
7 o C
14 o C
21 o C
28 o C
Leaf R mass
(nmol CO 2
g -1 s -1 )
10
5
100
50
b
d
10
5
10 50 100 150
LMA (g m -2 )
1 5 10
Leaf N concentration (%)
Acclimation adjusts the
LMA:R intercept, but not
the slope, so is
predictable from the
existing rate of
respiration and the
temperature change
from the reference value

Inclusion of temperature
acclimation in JULES
We assume that the ‘reference temperature’ for respiration is 25 o C
18
Respiraton (arbitary units)
16
14
12
10
8
6
4
2
0
Low temperature
ecosystems
No acclimation
+ acclimation
R mT
Tropics &
hot-arid
ecosystems
0 5 10 15 20 25 30 35 40 45 50
Temperature ( o C)

1861 2100
Acclimation of respiration
Acclimation-dependent
change in plant R
(% control)
Latitude
Acclimation-dependent
change in plant R
(g C m -2 yr -1 )
Control rates of
plant R (no acclimation)
(g C m -2 yr -1 )

Conclusions
Expectation is that
acclimation reduces
respiration with
increasing temperature.
Our net result is no
change in carbon balance
as +ve+
acclimation in
boreal zone cancels out -
ve acclimation in tropical
zone