Building tree-ring chronologies and Program ARSTAN
19 likes6,831 views
The document discusses dendroclimatology and paleoclimatology, including how tree rings can be used as climate proxies to reconstruct past climate variations. It explains how tree-ring measurements are standardized and aggregated to extract climate signals and maximize the environmental information while minimizing noise. Methods like COFECHA, detrending, and ARSTAN are used to process tree-ring data and develop high-quality tree-ring chronologies suitable for climate reconstruction.
![“ The main purpose of Program COFECHA
is the identification of [tree-ring] data
that should be reexamined for possible error.
”
Richard Holmes
Tree-Ring Bulletin 1983](https://image.slidesharecdn.com/thursday-120602155856-phpapp02/85/Building-tree-ring-chronologies-and-Program-ARSTAN-54-320.jpg)
Building tree-ring chronologies and Program ARSTAN
- 1. READING Dendroclimatology Brian Luckman Encyclopedia of Quaternary Sciences, 2006
- 2. READING A conceptual linear aggregate model for tree rings Ed Cook Methods of Dendrochronology, 1990
- 4. “ We’re all restless teens when some authority figure with an overhead projector starts yammering. ” Noel Murray The A.V. Club
- 6. observation
- 8. The Central England Temperature record is the longest instrumental record of temperature in the world. Degrees celsius relative to the long-term mean Source: Parker et al., Journal of Climate, 1992
- 9. Source: Goddard Institute for Space Studies, NASA
- 10. The number of climate stations recording air temperature falls off rapidly prior to AD 1900. Number of stations in the Northern Hemisphere Source: Jones et al., Journal of Geophysical Research, 2012
- 11. anecdote < observations <<
- 12. CLIMATE HISTORY OF NORTH AMERICA Younger Demise of Laurentide Dryas Ice Sheet 20 16 12 8 4 0 THOUSANDS OF YEARS AGO Final Drainage of Lake Agassiz LAST GLACIAL MODERN MAXIMUM OBSERVATIONS
- 14. CLIMATE PROXIES ice cores tree rings lake sediments speleothems corals
- 15. PALEOCLIMATOLOGY the study of climate prior to the period of instrumental measurement
- 16. Source: Tim Shanahan, University of Texas at Austin
- 17. Source: Geological Survey of Canada
- 18. Source: LACCORE, University of Minnesota
- 19. OBSERVATIONS PROXIES MODELS
- 21. READING Holocene perspectives on future climate change Ray Bradley Natural Climate Variability and Global Warming, 2008
- 24. “ Tree-ring analysis is one of the most powerful tools available for the study of environmental change and the identification of fundamental relationships ” between tree growth and climate. Ed Cook and Neil Pederson Lamont-Doherty Earth Observatory
- 25. “ RINGS ” IN THE BRANCHES OF SAWED TREES SHOW THE NUMBER OF YEARS AND, ACCORDING TO THEIR THICKNESS, THE YEARS WHICH WERE MORE OR LESS DRY. Leonardo da Vinci
- 26. “ The trees composing the forest rejoice and lament with its successes and failures and carry year by year something of its story in ” their annual rings. A. E. Douglass University of Arizona
- 27. 755 m3/s 847 m 3/s 809 m 3/s 770 m 3/s 823 m 3/s 787 m 3/s 901 m3/s 3
- 31. ACTIVITY TIMELINE Understanding COFECHA
- 32. ACTIVITY TIMELINE The physical basis for dendroclimatology
- 33. ACTIVITY TIMELINE Using ARSTAN
- 34. ACTIVITY TIMELINE Assessing chronology quality
- 39. ?1 Where (and how) can we obtain tree-ring data?
- 43. DOWNLOAD THESE DATA Bear Canyon West (NM586) Rito de los Frijoles (NM501) Fenton Lake Recollection (NM587) Baca (NM558) Los Alamos (NM044) Abouselman Spring (NM555) Los Alamos (NM046) Cat Mesa (NM556)
- 49. ?2 How can we tell if publicly-available ringwidth measurements are good quality?
- 53. COFECHA
- 54. “ The main purpose of Program COFECHA is the identification of [tree-ring] data that should be reexamined for possible error. ” Richard Holmes Tree-Ring Bulletin 1983
- 55. READING Computer-assisted quality control in tree-ring dating and measurement Richard Holmes Tree-Ring Bulletin, 1983
- 67. EXERCISE Use COFECHA to confirm that our sets of ringwidth measurements are dated correctly.
- 68. QUALITY CONTROL THESE DATA Bear Canyon West (NM586) Rito de los Frijoles (NM501) Fenton Lake Recollection (NM587) Baca (NM558) Los Alamos (NM044) Abouselman Spring (NM555) Los Alamos (NM046) Cat Mesa (NM556)
- 70. 755 m3/s 847 m 3/s 809 m 3/s 770 m 3/s 823 m 3/s 787 m 3/s 901 m3/s 3
- 71. ?3 Why should tree-ring variables be connected to climate?
- 74. Pinus spp. Source: Paul Schulte
- 75. TEMPERATURE high growth low growth cold hot
- 76. TEMPERATURE high growth frozen water low photosynthetic rate shorter growing season low growth cold hot
- 77. TEMPERATURE high growth low photosynthetic rate higher evaporation low growth cold hot
- 78. “ The growth of trees is undoubtably controlled more by the movement of water than by the movement of any other single substance.” Hal Fri s Tree Rings and Climate
- 79. WATER high growth low growth dry wet
- 80. WATER high growth reduced cell division reduced cell expansion C02 starvation low growth dry wet
- 81. WATER high growth flooding anoxic conditions low growth dry wet
- 83. Climate acts to synchronize growth rates at the level of the cell, the tree, the forest and beyond.
- 85. READING How well understood are the processes that create dendroclimatic records? Eugene Vaganov, Kevin Anchukaitis and Michael Evans Dendroclimatology, 2012
- 86. EARLYWOOD AND LATEWOOD WIDTH MAXIMUM LATEWOOD DENSITY VESSEL SIZE AND DISTRIBUTION WOOD BIOCHEMISTRY STABLE ISOTOPES TOTAL RING WIDTH
- 87. Tree-ring display at elementary school Photograph:Tom Swetnam
- 88. ?4 How do we extract climate information from a (complex) set of tree-ring measurements?
- 89. TREE-RING WIDTH DATA pith bark
- 90. “ THERMOMETERS ” TREES ARE NOT OR RAIN GAUGES. Keith Briffa and colleagues
- 91. Ed Cook Lamont-Doherty Earth Observatory
- 92. THE PRINCIPLE OF AGGREGATE TREE GROWTH
- 93. READING The decomposition of tree-ring series for environmental studies Ed Cook Tree-Ring Bulletin, 1987
- 94. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et
- 95. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et disturbance within the forest
- 97. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et disturbance from outside the forest
- 99. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et random processes not accounted by other sources
- 100. SIGNAL vs. NOISE
- 101. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et
- 102. ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic ation replication replication replication replic
- 103. never trust one tree
- 104. THE PRINCIPLE OF REPLICATION Making measurements from (i) more than one radius per tree and (ii) more than one tree per site maximizes the environmental signal and minimizes the amount of environmental ‘noise’.
- 105. Et can be assumed to be uncorrelated within and between trees in a stand.
- 106. δD1t Endogenous (‘originating within’) disturbances will be random events in both space and time, if the stand of trees is large enough.
- 107. δD2t o en is shared by most or all trees within a stand, but may not be shared by all forest stands within a region.
- 108. THE PRINCIPLE OF AGGREGATE TREE GROWTH Rt = At + Ct + δD1t + δD2t + Et size-related growth trend caused by physiological aging
- 109. Ring width Tree age
- 111. At does not have a universal or predictable shape.
- 112. “ At should be thought of as a nonstationary, stochastic process which may, as a special case, be modeled as a determinstic process. ” Ed Cook Tree-Ring Bulletin 1987
- 113. STANDARDIZATION
- 114. READING Standardization of tree-ring data Ray Bradley Paleoclimatology, 1999
- 115. “ Growth functions are removed by fi ing a curve to the data and dividing each measured ring-width value by the "expected" value on the growth curve. ” Ray Bradley Paleoclimatology 1999
- 116. (A) the ‘raw’ ring-width data
- 117. (B) the ‘detrending’ curve
- 118. DIVIDE (A) the raw ring-width data by (B) the ‘detrending’ curve
- 119. the ‘detrended’ ring-width index
- 120. “ Standardization transforms the non-stationary ring-widths in a new series of stationary, relative tree-ring indices that have a defined mean of 1.0 and a constant variance. ” Ed Cook Tree-Ring Bulletin 1987
- 121. ARSTAN autoregressive standardization
- 123. PROGRAM ARSTAN PRODUCES CHRONOLOGIES FROM TREE-RING MEASUREMENT SERIES BY DETRENDING AND INDEXING (STANDARDIZING) THE SERIES, THEN APPLYING A ROBUST ESTIMATION OF THE MEAN VALUE FUNCTION TO REMOVE EFFECTS OF ENDOGENOUS STAND DISTURBANCES.
- 124. DIVIDE (A) the raw ring-width data by (B) the ‘detrending’ curve
- 130. ARSTAN OUTPUT
- 133. CHRONOLOGY
- 136. EXERCISE Use several different detrending methods to standardize our set of ring-width chronologies and let’s see what happens!
- 137. STANDARDIZE THESE DATA Bear Canyon West (NM586) Rito de los Frijoles (NM501) Fenton Lake Recollection (NM587) Baca (NM558) Los Alamos (NM044) Abouselman Spring (NM555) Los Alamos (NM046) Cat Mesa (NM556)
- 138. USE FOUR DIFFERENT APPROACHES TO STANDARDIZATION Horizontal line Negative exponential curve 20-yr flexible spline Flexible spline set to 67% of series length
- 140. How should we plot our results?
- 141. DPL : YUX
- 142. WHO WILL FIND THE FIRST “DIRTY DOG”? Source: Terry Bain
- 143. USE FOUR DIFFERENT APPROACHES TO STANDARDIZATION Horizontal line Negative exponential curve 20-yr flexible spline Flexible spline set to 67% of series length
- 144. Pinus longaeva in western North America show a recent increase in growth even without any standardization to remove age-effects. Source: Saltzer et al., Proceedings of the National Academy of Sciences, 2009
- 145. READING Recent unprecedented tree-ring growth in bristlecone pine at the highest elevations and possible causes Ma hew Salzer, Malcolm Hughes, Andy Bunn and Kurt Kipfmeuller Proceedings of the National Academy of Sciences, 2009
- 148. READING The ‘segment-length curse’ in long tree-ring chronology development for palaeoclimatic reconstruction Ed Cook, Keith Briffa, David Meko, Donald Graybill and Gary Funkhouser The Holocene, 1995
- 149. The maximum wavelength of recoverable climatic information is related to the lengths of the individual tree-ring series used to construct the millennia-long chronology.
- 150. LOWEST RESOLVABLE FREQUENCY 3/n when n represents the average lengths of segments that make up the chronology
- 151. the ‘detrended’ ring-width index
- 152. Autocorrelation describes the correlation of a time series with its own past and future values.
- 153. Ringwidth lag-0 autocorrelation 3 2 1 0 -1 -2 -3 1900 1920 1940 1960 1980 2000 Year (A.D.)
- 154. Ringwidth lag-1 autocorrelation 3 2 1 0 -1 -2 -3 1900 1920 1940 1960 1980 2000 Year (A.D.)
- 155. Ringwidth lag-2 autocorrelation 3 2 1 0 -1 -2 -3 1900 1920 1940 1960 1980 2000 Year (A.D.)
- 156. Ringwidth lag-3 autocorrelation 3 2 1 0 -1 -2 -3 1900 1920 1940 1960 1980 2000 Year (A.D.)
- 157. covariance product of the standard deviation Autocorrelation
- 159. The spectrum of a time series describes the distribution of variance of the series as a function of frequency.
- 160. yellow violet blue green orange red
- 161. yellow violet blue green orange red short long wavelengths wavelengths
- 162. yellow violet blue green orange red Fast slow changes changes
- 163. 4 example of a ‘white’ time series 3 2 1 0 -1 -2 -3 200 250 300 350 400 450 500
- 164. 4 example of a ‘red’ time series 3 2 1 0 -1 -2 -3 200 250 300 350 400 450 500
- 165. 4 example of a ‘blue’ time series 3 2 1 0 -1 -2 -3 200 250 300 350 400 450 500
- 166. “White”
- 167. “Red”
- 168. “Blue”
- 170. IF a time series (of length N) is significantly autocorrelated, then: The series is not random in time Each observation is not independent from other observations The number of independant observations is fewer than N
- 171. STANDARD RESIDUAL
- 172. CAT MESA, NEW MEXICO ‘Standard’ chronology ‘Residual’ chronology
- 173. ?YEAH, BUT WHICH CHRONOLOGY SHOULD I USE?
- 174. PDO index Mexican PDSI 3 10 2 5 1 0 0 -1 -5 -2 -3 -10 1900 1920 1940 1960 1980 2000
- 175. The “effective sample size” is an estimate of the “real” number of observations a er adjusting for the effects of autocorrelation.
- 176. sample size effective sample size first-order autocorrelation Effective sample size
- 177. ?5 How can we determine if the common signal embedded within a set of tree-ring measurements is strong or weak?
- 179. ? To what degree does the chronology represent the hypothetical population chronology?
- 181. r bt The mean inter-series correlation calculated between all possible pairs of indexed series drawn from different trees.
- 182. How self-similar is this set of detrended ring-width measurements?
- 183. How self-similar is this set of detrended ring-width measurements? r = 0.51 bt
- 184. EPS quantifies the degree to which this particular chronology portrays the hypothetically perfect chronology.
- 185. EPS = f(rbt, t)
- 187. CAT MESA, NEW MEXICO ‘Standard’ chronology ‘Residual’ chronology
- 188. READING On the average value of correlated time series, with applications in dendroclimatology and hydrometeorology Tom Wigley, Keith Briffa and Phil Jones Journal of Climate and Applied Meteorology, 1984
- 192. READING Millennial precipitation reconstruction for the Jemez Mountains, New Mexico, reveals changing drought signal Ramzi Touchan, Connie Woodhouse, Dave Meko and Craig Allen International Journal of Climatology, 2011
- 193. PRODUCE CHRONOLOGIES FOR THESE DATA Bear Canyon West (NM586) Rito de los Frijoles (NM501) Fenton Lake Recollection (NM587) Baca (NM558) Los Alamos (NM044) Abouselman Spring (NM555) Los Alamos (NM046) Cat Mesa (NM556)