temporal framework for monitoring rangeland sustainability: a discussion by robert a....
TRANSCRIPT
Temporal Framework for MonitoringRangeland Sustainability: A Discussion
By
Robert A. Washington-AllenResearch & Development Staff
Environmental Sciences DivisionOak Ridge National Laboratory
P.O. Box 2008, MS 6407Oak Ridge, TN 37831-6407
E-mail: [email protected]
What is a temporal framework?
What is monitoring?
What is condition?
What is trend?
What is a standard or reference?
When do you begin monitoring?
How do you relate indicators?
What about causality?
Monitoring (Applied Ecology) Objectives
1. Detect change in the extent and distribution of indicators.
2. Assess rangeland condition and trend.
3. Suggest causality
4. Assess the risk of future crises.
O’Neill, R.V., C. Hunsaker, and D. Levine. 1994. Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.
Monitoring Challenges:
•Landscapes are middle-number systems.
•When indicators are re-measured some values will have changed:
•Does the change indicate a trend or normal fluctuations (historical variation)?
•Reliable evidence of trends require monitoring over long periods.
•For rangelands, current literature suggests 15 to 20 years if you wish to capture periodic climatic events (e.g., ENSO, La Nina, and PDO) or fire return intervals.
From: Turner, M.G., R.H. Gardner, and R.V. O’Neill. 2001. Landscape ecology in theory and practice. Springer-Verlag, New York, N.Y.
Illustration of Ecological Statics and Dynamics
The relationship of space and time to observable phenomena. The paradigm suggests the scales at which phenomena will act, provides an investigator with the constraints on experimental design, and suggests the tools required to detect phenomena of interest (adapted from Graetz 1987).
Landsat
Ground
domain of pastureproduction
domain oflandscape
degradation
0.10
1
10
100
1000
10000
100000
individual plants
water erosion features
Landsat MSS pixel sizesheet erosion
Landsat TM pixel size
Land SystemsCatena
Size of ranch
Ecological sites/land units
Foraging range of sheep
Foraging range of cattle
1
Temporal; Scale (days)
10000010000100010010
Spatial scale (m)
Rai
nfal
l eve
nt
Veg
etat
ion
gre
enup
Landsat
overpass
Lifetime of
most perennial
plants
Estimated
frequency
of severe
droughts
Estimated
frequency
Of wilfires
in Sagebrush
steppe
Soil
regeneration
A principle of Ecosystem science and Landscape Ecology is that before any ecosystem or landscape can be studied it must be bounded in space and time.
Grain and Extent must be explicitly stated.
Grain constrains the spatial and temporal scale of observation.
The coarsest resolution defines the studies grain.
NALC MSS 58m
MSS 79m
Condition: A one-time measure of change in an indicator relative to a standard.
Trend: A temporal measure of change in an indicator. The magnitude of a trend can be measured using a significant regression coefficient of the indicator versus time.
Standard: The reference that provides the basis for comparison that allows determination of a significant change. Reference conditions can be subjective or objective: the average (mean, mode, or median), maximum, or minimum conditions.
Ways to Monitor
Levels of Control Differ
Experimental:LaboratoryLab-FieldField
Retrospective Studies: Opportunistic, but usually lack controls and replications. Studies of longer temporal scales depend upon conditions which may no longer exist, i.e., suffer from lack of analogues for contemporary comparisons.
Computer Simulations
Also Note: Regional scale studies: Usually lack controls and replication, consequently not amenable to traditional statistical designs.
However, the systematic determination of a statistically significant trend (or an ecologically significant trend)may take years to determine.
O’Neill, R.V., C. Hunsaker, and D. Levine. 1994. Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.
Time series analyses set the analytical domain for trend data. This includes:
Data transformations:LogMoving averageDifferencing
Analystics:AutocorrelationAutoregressionNon-linear time series Spectral/FourierCross correlationIntervention
How do these metrics relate to each other and through time?
The example above can be viewed either as 3 different sites being compared or 1 site moving through its multivariate envelope.
Historical Variationatt
rib
ute
time
att
rib
ute
time
CHRONIC DISTURBANCE ACUTE DISTURBANCE
alternative
stateend
start
disturbance
malle
abili
ty
am
plit
ude
hysteresis
threshold
Adapted from Westman (1988), Hosten (1995).
The dry season soil-adjusted vegetation index (SAVI) statistical phase portrait for the sagebrush steppe portion of Deseret Land & Livestock Company ranch from 1972 to 1997.
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
SAVI Mean
SAVI Variance
Increasing Heterogeneity of Greenness
Increasing Greenness
9772
95
84
82
88
74
86
76
81
96
87
79 75
9091
7385
8980
Mean VarianceGrand MeanCentroidCluster
Legend
1994 1995 1996 1997
Dry Season
1980 1981 1982 1984 1985 1986
1987 1988 1989 1990 1991 1992
The dry season time series of soil-adjusted vegetation index (SAVI) images of the sagebrush steppe portion of Deseret Land & Livestock Company ranch from 1972 to 1997.
1972 1973 1974 1975 1976 1979
T4,T6
T4
Shrubland (B) Native Grassland (G)
T3 T3
T3,T5 Introduced Grasses (F)> 60 % Shrub Dense Shrubland (D)
T1: fire, T2: grazing, T3: heavy grazing, T4: cultural inputs, T5: drought, T6: wetter than average years
Threshold
Threshold
T5
T2
T1,T6
shrub grass/bare soil
sparse grass/bare soil
dense grass/bare soil
denser grass/bare soil
Spatial hypothesis of temporal change in landscape pattern and structure in relation to climate change, grazing, and fire. The spatial map was adapted from West and Young (2000).
1972
1973
1974
1975
1976
1979
1980
1981
1982
1985
1986
1987
1988
1989
1990
1994
1991
1992
1995
1996
1997
The growth form and land cover (bare ground) time series of thematic maps of the sagebrush steppe dominated portion of Deseret Land & Livestock Co. Ranch. The maps were derived from dry season Landsat MSS and TM satellite imagery from 1972 to 1997.
Time
% Grass Cover
% Shrub Cover
Contagion
The predicted response of growth-form composition and contagion to relatively high frequency of droughts and high grazing intensity. These predictions are the results of 200-year computer simulation model developed by Li and Reynolds (1998). The rate of change in contagion and physiognomy happened rapidly and abruptly relative to low and moderate frequencies of drought and grazing pressure.
Grass CoverShrub Cover
Dry Season Year
Shrub/Grass Cover Ratio
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Grand Mean
Mean 1972 - 1987
Mean 1987 - 1997
Sequence of Years
3020100
54
52
50
48
46
44
42
r2 = 0.35p = 0.0015n = 26
Lag Year
21191715131197531
1.0
.5
0.0
-.5
-1.0
•Monitoring can suggest, but it can seldom demonstrate causality.
•Monitoring can only hope to show correlations.
O’Neill, R.V., C. Hunsaker, and D. Levine. 1994. Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.
Garff, Freed, & Robinson 1972-1975
1972
1973
1973
1974
1974
1975
1975
Season-Year
Mean SAVI
J. Hotung 1975 -1983
0.200.220.240.260.280.300.320.340.360.380.400.420.440.460.480.500.520.540.560.580.60
1975
1975
1976
1976
1977
1977
1978
1978
1979
1979
1980
1980
1981
1981
1982
1982
1983
1983
Season-Year
Mean SAVI
LDS Farm Management Group 1983 - 1998
Season-Year
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
Season-Year
Mean SAVI
Mean
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
1953-1975 1975-1983 1983-1998
ManagementPeriod
Cubic Fit
The seasonal wet and dry soil-adjusted vegetation index (SAVI) time series for the sagebrush steppe portion of Deseret Land & Livestock Company ranch from 1972 to 1998. The 3 management regimes and the best fit for each regime and the entire time series are curvilinear regression line are delineated. Missing years were replaced by linear interpolation.
Y= 0.02x + 0.29r2 = 0.32p = 0.18n = 7
Y= 0.0005x2 - 0.01x + 0.41r2 = 0.07p = 0.58n = 16
1983
1983
1984
1984
1985
1985
1986
1986
1987
1987
1988
1988
1989
1989
1990
1990
1991
1991
1992
1992
1993
1993
1994
1994
1995
1995
1996
1996
1997
1997
1998
Mean SAVI
Y= 0.0005x2 - 0.01x + 0.46r2 = 0.40p = 0.003n = 31
0.200.220.240.260.280.300.320.340.360.380.400.420.440.460.480.500.520.540.560.580.60
0.200.220.240.260.280.300.320.340.360.380.400.420.440.460.480.500.520.540.560.580.60
(cubic fit, r2 = 0.29, p = .003)
Continuous Grazing Rotational Grazing Short Duration Grazing
Mean interannual Palmer Drought Severity Index (PDSI) from 1895 to 1996 for the Northern Mountains Climatic Region 5. PDSI values from -4 or less indicate extreme drought and from 4 or greater extreme wet periods (Alley 1984). These data were acquired from the Utah Climate Center.
Palmer Drought Index for Utah Region 5 from 1895 - 1997
-8
-6
-4
-2
0
2
4
6
8
18951898190119041907191019131916191919221925192819311934193719401943194619491952195519581961196419671970197319761979198219851988199119941997
Year
PDSI
Dust Bowl 1950s 87 to 89
1986 1987 1988 1989 1990 1991
1980 1981 1982 1983 1984 1985
1992 1993 1994 1995 1996 1997The apparent spatial distribution of livestock from 1980 to 1997 at the paddock-level on Deseret Land & Livestock Company Ranch.
, North Dakota
http://www.ngdc.noaa.gov/paleo/drought/drght_2000years.htmlLaird, K. R., S. C. Fritz, K. A. Maasch, and B. F. Cumming. 1996. Greater drought intensity and frequency before A.D. 1200 in the Northern Great Plains, U.S.A. Nature 384:552-554.
1200
Drought in the last 2000 years
Hughes, M. K. and L. J. Graumlich. 1996. Climatic variations and forcing mechanisms of the last 2000 years. Volume 141. Multi-millenial dendroclimatic studies from the western United States. NATO ASI Series, pp. 109-124.
Landscapes are middle-number systems.
Farming is still a high-risk venture.
Fires (yellow polygons) detected on the Deseret Land & Livestock Company Ranch from 1992 to 1996 using Landsat Multispectral Scanner and Thematic Mapper (TM) satellite imagery. An 1994 Indian Resource Satellite (IRS) scene, which has been merged to a false-color TM scene, and a polygon coverage of the grazing paddocks serve as the backdrop.
1992
1994
1994
1994
1996
Retrospective inference of the relationship of landscape-scale metrics with extent and perimeter of fire on the eastern portion of Deseret Land & Livestock Co. Ranch from 1972 to 1997 using linear regression. For the relationship of fire perimeter with landscape-scale metrics, only the significant correlation of with contagion ( p ≤ 0.10) is shown.
Perimeter of Fire (m)
40000300002000010000
47
46
46
45
45
44
44
43
43
1996
1994
1992
Extent of Fire (ha)
2000180016001400120010008006004002000
46.5
46.0
45.5
45.0
44.5
44.0
43.5
43.0
42.5
1996
1994
1992
Extent of Fire (ha)
2000180016001400120010008006004002000
60.5
60.0
59.5
59.0
58.5
58.0
57.5
57.0
56.5
1996 1994
1992
Extent of Fire (ha)
2000180016001400120010008006004002000
960
940
920
900
880
860
1996
1994
1992
Number of Patches Mean Patch Size Contagion
ContagionNearest Neighbor SDEMean Nearest Neighbor
r = -0.99931p = 0.02n = 3
r = 0.99p = 0.09n = 3
r = -0.32p = 0.79n = 3
r = -0.96p = 0.17n = 3
r = -0.92p = 0.26n = 3
r = -0.99p = 0.07n = 3
0
2000
4000
6000
8000
10000
12000
14000
16000
1891
1895
1899
1903
1907
1911
1915
1919
1923
1927
1931
1935
1939
1943
1947
1951
1955
1959
1963
1967
1971
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Year
Animal Unit (AU)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
Mean SAVI
Grand Mean = 7034 AU1891 to 1903 = 7548 AU1915 to 1951 = 10633 AU1967 to 1998 = 4444 AU
The time series of seasonal soil-adjusted vegetation index (SAVI) from 1972 to 1998 (line plot) in relation to the time series of animal units (combined sheep and cattle, bar chart) from 1891 to 1998.
A comparison between the 1972 to 1997 time series of the seasonal month of image acquisition Palmer Drought Severity Index (PDSI), the 5-year moving average of PDSI, and the seasonal soil-adjusted vegetation index (SAVI) of the Rich County-sagebrush steppe portion of Deseret Land & Livestock Company ranch.
0.00
0.10
0.20
0.30
0.40
0.50
0.60
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
wet
dry
Season
Mean SAVI
-8
-6
-4
-2
0
2
4
6
8
Year
PDSI
Mean SAVI PDSI 5 per. Mov. Avg. (PDSI)
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
NPy = -793x + 1052
R2 = 0.81p =0.0004
n = 10
MPSy = 64x + 48
R2 = 0.89p = 0.00004
n = 10
600
650
700
750
800
850
900
950
1000
1050
1100
0.000 0.100 0.200 0.300 0.400 0.500 0.600
Bulk Stocking Density (AU/Ha)
Number of Patches
0
10
20
30
40
50
60
70
80
90
Mean Patch Size (Ha)
Land NP MPS Linear (Land NP) Linear (MPS)
Grazing Pressure
Soil MoistureEl Niño
La Niña
Shrubland
Grassland
A
Low Production
High Production
B
C D
E
F
Erosion
Deposition
Clumped
Fragmented
Modified to 3-D from Holmgren et al. (2001)
Vostok Ice Core: The Vostok temperature record indicates that the earth has been colder than present for most of the past 250,000 years, including many ice ages .
J. Jouzel, C. Waelbroeck, B. Malaiz, M. Bender, J. R. Petit, N. I. Barkov, J.M. Barnola, T. King, V. M. Kotlyakov, V. Lipenkov, C. Lorius, D. Raynaud, C.Ritz and T. Sowers, Climatic interpretation of the recently extended Vostok ice records, Clim.Dyn., In press
Years before present (1000)
Temp.
How long do you monitor?
Longer time scales: Robert Frost Planning Horizon
Fire and Ice