| Eleven new Pinus echinata Mill. site chronologies from the Ouachita Mountains are created and combined into a single master chronology. Stands used are naturally-regenerated even-aged stands from 31 to 229 years old. On high-quality sites, missing rings occur only in severe storm years. The Expressed Population Signal is greater than 0.85 for the years 1783 to 2009. Sample size exceeds thirteen trees from 1872 to 2009. It is suitable for use in dendroarcheology, dendroecology, climate change and weather studies. There is a listing of the master and component site chronologies in Tucson format. Historical records of severe winter storms only extend back to 1892; dendrochronology extends this to 1745. The winter storm signal is consistent with injury caused by trunk breakage, branch loss and severe bending. This suggests greater exposure by large crowned trees. Drought may be associated with winter storms; use of Palmer's Drought Severity Index to detrend data risks removal of the severe storm signal. Cohen's Kappa is used to test four different methods of storm detection in tree rings. Average storm recurrence interval is 17 years; two out of three known ice storms produce trunk breakage. Severe winter storm reconstructions are presented for thirteen study locations. Many variables affect tree breakage and height loss. Multiple linear regression and logistic models use total tree height, diameter and live crown ratio to predict the probability of trunk breakage and height loss. Total height, diameter and live crown ratio account for 22.3% of variation; height alone accounts for 15.4%. Logistic brekage probability models have p-values of 0.0001 (height) and 0.0277 (diameter). Further research might permit the separation of wind and ice storm signals and a means of estimating second and third-quarter precipitation and drought. Tree ring chronologies are a powerful tool for weather and climate studies at a finer scale than is possible with other proxy data. |
