Long-term variability of the Land Surface

1
Long-term variability of the Land Surface
Dynamic vegetation

• Lecture 13
• CLIM 714
• Paul Dirmeyer

2
Time scales of variability
Locally, any land surface state variable varies
on a range of time scales
3
Milankovitch Cycles
Today the Earth experiences about a 6 difference
in the amount of solar radiation received in
January compared to July. When the Earth's orbit
is more elliptical, the amount of energy received
would be vary much more between seasons, in the
range of 20-30.
This inclination oscillates in a range of 21.8o
and 24.4o.
Precession oscillates between the two positions
in a period of about 22,000 years. (The 22 000
year cycle is in fact a combination of a 19 000,
and a 23 000, year cycle).
The three cycles combine to produce variations in
the amount of heating and the length of the
seasons. The effect is most pronounced when the
Earth is farthest from the Sun during the
northern winter. The northern hemisphere is
critical to the formation of large glaciers
because most of the land is concentrated there.
The glaciers grow not because of overall
temperature decreases, but because there is not
enough heating during the summer to melt the
accumulated ice.
4
Long term change - past
5
Todays vegetation
6
8000 ybp
7
11,000 ybp
8
13,000 ybp
9
18,000 ybp
10
Europe Present Potential Vegetation
11
Europe Past vegetation
12
Africa Current potential vegetation
13
Past African Climate
14
Historical Climate Change in the Mediterranean
Basin the Role of Vegetation Feedbacks
The Motivation Northern Africa in Roman Times was
considered one of the most prosperous and rich
areas of the western world. The production of
wheat, olive oil, and wine was greater than in
Mediterranean Europe. Strabo (I century C.E.) and
other classical authors describe northern Africa
as a strip of vegetated land, and place the
northern border of the desert a several hundred
kilometers to the south of the sea. In present
times, most of the coastal regions of northern
Africa are sub-desert or desert. Pliny (I century
C.E.) speaks of elephants living to the south of
the Atlas range, in an area that is now desert
Ptolemy (II century C.E.) describes summer
thunderstorms in Alexandria, where now summers
are dry. Archeology confirms the presence of
heavy agricultural activity in areas which are
now classified as hyper-arid. Modern palynology
has provided evidence of a trend towards drier
conditions throughout the entire Mediterranean
region.
In the numerical experiment, modern vegetation
was replaced by forests and grasslands in the
hatched areas, consistent with evidence from the
Roman Classical Period (ca. 2000 y.b.p.)
Reale Dirmeyer Reale Shukla, (2000 Glob.
Planet. Change)
15
Historical Climate Change in the Mediterranean
Basin the Role of Vegetation Feedbacks
The model experiment A sensitivity test with a
low-resolution general circulation model reveals
that the position of the Inter-Tropical
Convergence Zone (ITCZ) is sensitive to changes
in surface properties far to the north, around
the Mediterranean region (Reale and Dirmeyer,
Glob. Plan. Ch. 2000). Reale and Shukla, (Glob.
Plan. Change, 2000) quantify the response of a
climate model with respect to the land surface
conditions of two millennia ago. These are
inferred mostly from palynological studies. The
experiments indicate a change in the general
circulation of the atmosphere namely a northward
shift of the ITCZ over eastern Africa, and a
land-sea circulation similar to a small-scale
monsoon occurring between the Atlas range and the
Mediterranean. These changes benefit, with a
significant increase in precipitation, the Nile
Valley and the Atlas range, which were two of the
most agriculturally productive regions of the
Roman world. Thus, the experiment suggests that
the large-scale clearings that occurred in the
Late Antiquity and during the Middle Ages, may
have contributed to the dryness of the present
climate.
Simulated summer rainfall is substantially
greater over the Nile Valley and in the vicinity
of the Atlas Mountains when modern vegetation
distributions are replaced by the vegetation of
the Roman Classical Period.
Reale Dirmeyer Reale Shukla, (2000 Glob.
Planet. Change)
16
Long term change - Future
Fossil fuels Greenhouse gases Global
warming Anthropogenic climate change
17
Potential range changes of selected tree species
in Yellowstone region of the Rocky Mountains
under a projected climate based on a doubling of
atmospheric carbon dioxide.
18
A Climate Change Atlas for 80 Tree Species of
the Eastern United States Anantha M. Prasad and
Louis R. Iverson
predictions for Longleaf Pine
http//www.fs.fed.us/ne/delaware/atlas/
19
Has Vegetation Responded to Climate Change?
20
Continental Differences in Warming

• Overall warming in Eurasia. Less warming and even
  some cooling in North America

Land surface April-October temperature trends in
?C/18 yrs between 1982 and 1999 (NASA GISS
Station Temperature data)
21

• Study Region
• Vegetated pixels between 30?N-70?N

• Objectives
• minimize the effect of Solar Zenith Angle
• reduce background effects (snow, barren and
  sparsely vegetated areas)
• use data from the same pixels in the entire
  analysis.

22
Changes in Vegetation Activity

• Changes in vegetation activity can be
  characterized through
• changes in growing season
• changes in seasonal NDVI magnitude

Increases in NDVI magnitude
Increases in growing season
Increase
NDVI
23
Longer Growing Seasons
(Increased by 12 Days)
11.9 days/18 yrs (plt0.05)
(Increased by 18 Days)
17.5 days/18 yrs (plt0.05)
24
Increases in April-October NDVI Magnitudes
(8 Percent Increase)
8.4/18 yrs (plt0.05)
(12 Percent Increase)
12.4/18 yrs (plt0.05)
25

• Spatial Pattern of April-October NDVI Changes
• Persistence index an index for identifying
  regions where NDVI has increased consistently

A persistent increase in NDVI is observed in
Eurasia over a broad contiguous swath of land
while North America shows a fragmented pattern of
change.
26

• April-October NDVI trend at the 5 significance
  level

Pixels that show a statistically significant
trend are also pixels with high persistence.
27

• April-October NDVI difference between 1995-99 and
  1982-86 averages

Pixels that show a large NDVI increase are also
pixels with high persistence and statistically
significant trends.
28
Consistency between April-October NDVI and
Temperature  
R0.79 (plt0.01)
R0.72 (plt0.01)
Year-to-year changes in growing season NDVI are
tightly linked to year-to-year changes in
temperature.
29
Carbon Cycle - Current Uncertainties
5.50.3

• Current source
• and sink
• strengths are
• uncertain.
• Prediction of
• future climate
• forcing is
• therefore
• uncertain as
• well.

To Atmosphere
Unidentified 1.8 1.5 Sink
Atmospheric 3.3 0.2 Carbon
Land use 1.6 0.8 change
Ocean 2.0 0.6 Uptake
Fossil Fuels


-
-
To Land/Ocean
Atmospheric storage
human input
biosphere uptake
Peta (1015 ) grams of carbon/year
30
Can we predict vegetation changes?
31
(No Transcript)
32
Dynamic Vegetation Models (DVMs)
33
Elements to model.
The spatial and temporal dynamics of the
biosphere have strong implications for the
overall system
34
What drives long-term variability?
35
Succession models
36
The problem with succession models
They are linear.
37
DVMs and PFTs
38
Ecosystems are dynamic
39
In DVMs, autogenic succession is internal to the
modelClimate and humans are external
40
DVM example 1Boreal Forest
41
DVM Example 2Tropical Forest
42
Summary of the first phase of the PILPS C-1
project
Comparison of both  biophysical  and
 biogeochemical  flux from different types of
models with observations at one EUROFLUX site
Loobos, Netherlands

• The site
• Temperate mature (100 years) coniferous
  forest
• Climate 700 mm precipitation , 9.8 C mean
  temperature
• Planted on a sand ? no soil carbon at the
  beginning of the plantation
• Measured fluxes NEE, LE,H, Rn
• Meteorological parameters incoming SW rad.,
  precipitation, temperature, wind speed, relative
  humidity, pressure
• -Period covered 1997-1998

Models Including SVAT with and without carbon
cycle
www.pilpsc1.cnrs-gif.fr/

• Simulations
• Free equilibrium simulations
• Models are run until equilibrium of state
  variables using years 1997-1998  in loop 
• Free 100 years run
• simulation of  realistic scenario  Beginning
  with a soil with no carbon, the models are run
  for 1906 (plantation of the forest) to 1998 using
  observed climate.

43
DVM Comparison PILPS C1
Models were not calibrated in advance. This is
evident in the different trajectories of forest
growth and soil carbon storage during the
100-year period.
44
DVM simulation of global vegetation
One case NASA-CASA
45
DVM simulation of CO2 variability
46
DVM simulation of climate change response
Reversal in trend is due to the release of soil
carbon from high latitudes
47
Simulating disturbances
48
Dynamic vegetation modeling