Presentation is loading. Please wait.

Presentation is loading. Please wait.

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

Similar presentations


Presentation on theme: "Temporal Framework for Monitoring Rangeland Sustainability: A Discussion By Robert A. Washington-Allen Research & Development Staff Environmental Sciences."— Presentation transcript:

1 Temporal Framework for Monitoring Rangeland Sustainability: A Discussion By Robert A. Washington-Allen Research & Development Staff Environmental Sciences Division Oak Ridge National Laboratory P.O. Box 2008, MS 6407 Oak Ridge, TN

2 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?

3 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. ONeill, R.V., C. Hunsaker, and D. Levine Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.

4 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.

5 From: Turner, M.G., R.H. Gardner, and R.V. ONeill Landscape ecology in theory and practice. Springer-Verlag, New York, N.Y. Illustration of Ecological Statics and Dynamics

6 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 pasture production domain of landscape degradation individual plants water erosion features Landsat MSS pixel size sheet erosion Landsat TM pixel size Land Systems Catena Size of ranch Ecological sites/land units Foraging range of sheep Foraging range of cattle 1 Temporal; Scale (days) Spatial scale (m) Rainfall event Vegetation greenup Landsat overpass Lifetime of most perennial plants Estimated frequency of severe droughts Estimated frequency Of wilfires in Sagebrush steppe Soil regeneration

7 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.

8 Grain constrains the spatial and temporal scale of observation. The coarsest resolution defines the studies grain.

9 NALC MSS 58m MSS 79m

10 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.

11 Ways to Monitor Levels of Control Differ Experimental: Laboratory Lab-Field Field 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.

12 However, the systematic determination of a statistically significant trend (or an ecologically significant trend)may take years to determine. ONeill, R.V., C. Hunsaker, and D. Levine Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.

13 Time series analyses set the analytical domain for trend data. This includes: Data transformations: Log Moving average Differencing Analystics: Autocorrelation Autoregression Non-linear time series Spectral/Fourier Cross correlation Intervention

14 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.

15 Historical Variation attribute time attribute time CHRONIC DISTURBANCE ACUTE DISTURBANCE alternative state end start disturbance malleability amplitude hysteresis threshold Adapted from Westman (1988), Hosten (1995).

16 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 SAVI Mean SAVI Variance Increasing Heterogeneity of Greenness Increasing Greenness Mean Variance Grand Mean Centroid Cluster Legend

17 Dry Season 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

18 T4,T6 T4 Shrubland (B) Native Grassland (G) 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 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).

19 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.

20 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 Cover Shrub Cover

21 Dry Season Year Shrub/Grass Cover Ratio Grand Mean Mean Mean

22 r 2 = 0.35 p = n = 26 Lag Year

23 Monitoring can suggest, but it can seldom demonstrate causality. Monitoring can only hope to show correlations. ONeill, R.V., C. Hunsaker, and D. Levine Monitoring challenges and innovative ideas. Proc. Int. Sym. Ecological Indicators.

24 Garff, Freed, & Robinson Season-Year Mean SAVI J. Hotung Season-Year Mean SAVI LDS Farm Management Group Season-Year wetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywetdrywet Season-Year Mean SAVI Mean Management Period 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 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 r 2 = 0.32 p = 0.18 n = 7 Y= x x r 2 = 0.07 p = 0.58 n = Mean SAVI Y= x x r 2 = 0.40 p = n = (cubic fit, r 2 = 0.29, p =.003) Continuous GrazingRotational GrazingShort Duration Grazing

25 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. Dust Bowl1950s87 to 89

26 The apparent spatial distribution of livestock from 1980 to 1997 at the paddock-level on Deseret Land & Livestock Company Ranch.

27 , North Dakota Laird, K. R., S. C. Fritz, K. A. Maasch, and B. F. Cumming Greater drought intensity and frequency before A.D in the Northern Great Plains, U.S.A. Nature 384: Drought in the last 2000 years

28 Hughes, M. K. and L. J. Graumlich 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

29 Landscapes are middle-number systems. Farming is still a high-risk venture.

30 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

31 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. Number of Patches Mean Patch SizeContagion Nearest Neighbor SDEMean Nearest Neighbor r = p = 0.02 n = 3 r = 0.99 p = 0.09 n = 3 r = p = 0.79 n = 3 r = p = 0.17 n = 3 r = p = 0.26 n = 3 r = p = 0.07 n = 3

32 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.

33 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. wetdrywetdrywetdrywetdrywetdrywetdrywetdry wet dry Season Mean SAVI Year PDSI Mean SAVIPDSI5 per. Mov. Avg. (PDSI)

34 NP y = -793x R2 = 0.81 p = n = 10 MPS y = 64x + 48 R2 = 0.89 p = n = Bulk Stocking Density (AU/Ha) Number of Patches Mean Patch Size (Ha) Land NPMPSLinear (Land NP)Linear (MPS)

35 Grazing Pressure Soil Moisture El 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)

36 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?

37 Longer time scales: Robert Frost Planning Horizon Fire and Ice


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

Similar presentations


Ads by Google