Transcript Assessing the relative role of climate change and human activities in

Quantitative assessment of the relative role of climate change and human
activities in grassland degradation: Application of a satellite tracking system
FACULTY OF
AGRICULTURE
AND ENVIRONMENT
Inakwu O.A. Odeh
With Professor J Li and Team from Nanjing
University
Department of Environmental Sciences
Presentation for the Space SyReN (University of Sydney); November 18, 2014
Introduction
Grassland covers approximately 25% of world's natural land surface
It accounts for about 16% of the global terrestrial GNPP
Also, globally, grassland has a major influence on the functioning of the
terrestrial biosphere
2
Introduction
In China,
Grassland is one of the most important
natural resources
It accounts for 42% of the national land
area (and 11% of global grassland)
It is home to rich plant and animal
diversity
It is the major source of animal products
for the teeming population- products such
as meat, milk, wool and pelts
3
Introduction
However, grassland in China
experienced large-scale degradation
and desertification in the last 30-40
years due to:
Overgrazing
Large-scale conversion to croplands to feed the
teeming population
Drought
And suspiciously climate change
4
Introduction
In response, China introduced policies (late 1990s and early 2000s) to
restored degraded/ dysfunctional grasslands- extending to northwest
The restoration programs included
Three-North Shelterbelt Forest project,
The Grain-to-Green Project
 Grazing Withdrawal Project
5
Study Aim
› About 2010, a research project (Funded by Chinese Govt, AusAID, Asia‐Pacific
Network for Global Change Research and Usyd IPDF) was initiated
- to quantitatively assess the extent and degree of grassland degradation in response to
government restoration programs vis-à-vis the impact of climate change and variability on
grassland degradation
Grassland Types in Northwestern China
6
Project Team
› The project was carried out in collaboration with the University of Nanjing's Global
Change Institute (GCI-UN).
•
•
•
•
•
•
•
•
Professor Jianlong Li
Dr S Mu
Dr S Zhou
Dr C Gang
Dr W Ju
Y Chen
Dr Z Wang
Etc.
7
Methods
 The main thrust of the methodology used
was the ability to estimate Net Primary
Productivity (NPP) from satellite data and
using ground data for validation over such
a large region; Steps:
‘Actual’ NPP was estimated between 2001 and 2010 using CASA
(Carnegie-Ames-Stanford Approach) with MODIS NDVI as the
input data
Potential NPP was estimated using Thorntwaite Memorial model
based on meteorological data
Differences between potential and actual NPP are hypothesized to
be due to either climate change or human activities or both
8
Data Requirement
Data required and data processing
Meteorological data-
Including monthly mean temperature and precipitation,
total solar radiation were obtained from China Meteorological Data Sharing Service
System.
Land cover data: Global Land Cover 2000 dataset
Normalized difference vegetation index (NDVI) data (MODIS)-NDVI data
with 1 km spatial resolution from 2001 to 2010,
Field survey to estimate on-ground NPP-
We sampled 63 sites across the
study area in early April and at the end of August in 2009, to validate the accuracy of
the estimated NPP by model.
These datasets are processed within the ArcGIS10.1.
Methods- CASA Model for Computing Actual NPP
10
Methods- CASA Model for Computing Actual NPP
11
Methods- CASA Model for Computing Actual NPP
The light use efficiency  can be estimated as:
 x, t   T 1 x, t   T 2 x, t   W x, t    max
where
T 1 x, t  is a coeff.- represents the reduction of NPP caused by
biochemical action under extreme temperature conditions;
T 2 x, t  is a coefficient that determines the biomass decline
when the temperature deviates from the optimal temperature;
W x, t  is the moisture stress coefficient which is indicative of
the reduction of light-use efficiency caused by moisture factor;
 max is the maximal light-use efficiency under ideal conditions
= 0.542 for grasslands
12
Methods- CASA Model for Computing Actual NPP
13
NPP, and hence APAR, is a function of vegetation type and
vegetation cover- represented by vegetation indices
 In particular, a number of vegetation indices are products
of VIS-NIR (satellite) remote sensing systems, e.g.:
 Simple ratio (SR);
 Normalized difference vegetation index (NDVI)
 Fractional vegetation cover
14
Common types of vegetation indices
 The ratio of near-infrared (NIR) to red simple ratio (SR) is the first true
vegetation index:
red
NIR
NIR
SR 
red
 Takes advantage of the inverse relationship
between chlorophyll absorption of red radiant
energy and increased reflectance of near-infrared
energy for healthy plant canopies
Common types of vegetation indices
Normalized difference vegetation index (NDVI
red
NIR
NIR  red
NDVI 
NIR  red
 Used to
 identify ecoregions;
 monitor phenological patterns of the
earth’s vegetative surface, and
 assess the length of the growing season
and dry periods;
 estimate net primary production (NPP)
Fractional Vegetation Cover (fv)
 fv can be computed from NDVI by using a linear mix
model with two end members representing fully
vegetated land surface and bare ground:
Leaf Versus Canopy
0.5
very dense vegetation cover (Fv max)
Reflectance (%)
0.4
very scant Fv
sunlit bare soil
0.3
0.2
0.1
0.0
400
600
800
1000
Wavelength, nm
1200
FAPAR is a function of vegetation type and
vegetation cover
 FAPAR is a function of vegetation type and vegetation
cover;
 Vegetation type and cover can be modelled by satellite
remote sensing data, especially the visible/ near infrared
section of EM radiation;
 Satellite remote sensing is particularly advantageous
because of their archival databases that provide time
series records of the earth surface conditions
21
CASA Model for Computing Actual NPP
FPAR can be calculated from NDVI as:
FAPARNDVI 
( NDVI ( x ,t )  NDVI i ,min )  FAPARmax  FAPARmin 
( NDVI i ,max  NDVI i ,min )
 FPARmin
where
NDVImax and NDVImin are respectively 0.634, 0.023;
FAPARmax and FAPARmin are 0.95 and 0.001 respectively
22
Methods- Thornthwaite Memorial NPP Model for
Computing Potential NPP
23
Methods- Thornthwaite Memorial NPP Model for
Computing Potential NPP
Thornthwaite Memory model is expressed as:

NPP  3000 1  e
0.0009695v 20

where v is the average annual actual evapotranspiration (mm),
expressed as:
V
1.05r
1  1  1.05r L 
2
where L is annual average potential evapotranspiration (mm), expressed
as:
L  3000  25t  0.05t 3
and r is annual precipitation (mm), t is the annual average temperature
(℃)
24
Method- Computation of grassland vegetation
dynamics vis-à-vis roles of climate and humans
› Change trend of grassland NPP- whether actual or potential can be
obtained from the slope of NPP trend, S, calculated as a linear fit of
time/NPP using the ordinary least square estimation:
n
S
n
n
n   i  NPPi  ( i)(  NPPi )
i 1
i 1
i 1
n
n
i
i 1
n   i 2 （ i）2
› Significance test of change trend of grassland NPP can be done using
statistic F test.
where, U is regression sum of squares , Q is residual sum of
squares, n is the df = 9 years
Methods- Flowchart to determine relative roles of climate
change vs human activities to grassland dynamics
MODIS NDVI
data (20012010)
Weather station
data (2001-2010)
Human appropriation NPP
(NPPH) (NPPP -NPPA)
Actual NPP (NPPA)
from CASA model
Compute trend slope of NPPA
(SA) and ΔNPPA
Trend slope of NPPH (SH) and
ΔNPPH
Potential NPP (NPPP) from
Thorntwaite memorial model
Trend slope of NPPP
(SP) and ΔNPPP
Analyse relative roles of climate and
humans based on 8 scenarios
26
Methods- Scenarios of relative roles of climate change vs
human activities to restoration/degradation
ΔNPPj=(n-1)×Sj
27
Results: Spatial distribution of actual grassland
NPP in NW China (2010).
Actual grassland NPP in NW
China (2010).
28
Results- Validation of Estimated ‘Actual’ and
Potential NPP
Actual NPP
Potential NPP
1
The model accuracy of (a) CASA model (Actual NPP) and (b) Thornthwaite
Memorial model (Potential NPP)
29
Results: Grassland vegetation dynamics
（c）
Trending slope of NPP (grassland) dynamics
The proportion of different categories
of grassland dynamics
The degree of NPP dynamics
Result: Proportion of grassland
restoration/degradation by province
Area percentage of grassland degradation and restoration
31
Results: The relative roles of climate change
versus human activities on grassland degradation
climate change
human activities
The proportion of the relative roles of (a) climate change and (b) human activities to grassland
degradation.
32
Results: contribution of climate change/human activities to
grassland degradation/ restoration by province
Grassland degradation
Grassland restoration
Contribution of climate change, human activities and the combination of the two
factors to (a) grassland degradation; and (b) grassland restoration
33
Result: Spatial patterns of contributions of climate change
and human activities to grassland degradation
climate change
Human activities
Contributions of climate change (a) and human activities (b) to grassland restoration
34
35
Global extension- trend in grassland dynamics
36
Global extension- role of climate change vs human
activities to grassland degradation
37
Global extension- role of climate change vs human
activities to grassland restoration
38
39
Conclusions- NW China Study
The mean annual grassland NPP in 2010 was estimated to be about 123
g C/m2/yr and showed obvious spatial heterogeneity.
Between 2001-2010, 62% (1,650,316 km2) of total grassland was
degraded
Out of this, 66% of grassland degradation was caused by human activities
Only about 20% was due to climate change
Overall, 38% (1,033,663 km2) showed improvement
Satellite tracking can be useful for elucidating the performance of
grassland restoration programs through careful analysis
40
Conclusions
41
Pictures
42