Temporal Framework for Monitoring Rangeland Sustainability: A Discussion By Robert A. Washington-Allen Research & Development Staff Environmental Sciences.

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