The motion of trees in the wind - a collection of multiple data sets

The tree motion data in this repository were collected from multiple sites with different sensors. There are two types of data: (a) 1-hour sample data from 27 sites (/one_hour_samples/*.csv). Each file represents a single tree and contains two columns representing the two horizontal axes of tree motion. These data are uncalibrated and do not have a time-stamp. File names refer to: OwnersInitials_Site_treeID_resolution_SensorType for example TJ_MY_7_4Hz_Strain These filenames match the first column in the 'TreeSummary.csv' table (b) long-term data from 9 data sets (/site_name.zip/) The wind and tree data are stored in separate files for each day of data, labelled with site name, tree ID or wind, year, month, day. For example Kershope_T25_1990_5_23 In each file, the first column is the time and date in format YYYY-MM-DD HH:mm:ss.SSS The following columns are tree motion data or wind data with units in the variable names. Some of the data were calibrated by pulling the tree with a known force, in these cases calibration coefficients are given in the 'TreeSummary' table Wind data for the Australian data set is stored in the folder with data for each tree, because the trees were not close together enough to share a wind measurement. The data for Orange, Storrs and Torrington were collected with inclinometers. The data for all trees is stored together in each daily file. The long-term tree motion data have not been pre-processed because user choices affect the results. Importantly, you will need to remove a varying offset (caused by sensor drift) from the tree motion data. See discussion in the accompanying publication and references therein. References: Publication describing all collated data and necessary pre-processing. Jackson, T. et al (2021) The motion of trees in the wind: a data synthesis. Biogeosciences. Australian data decsribed in: James, K. et al (2006).: Mechanical stability of trees under dynamic loads, Am. J. Bot. doi:10.3732/ajb.93.10.1522 Kershope and Rivox data described in: Gardiner, B. A., et al. "Field and wind tunnel assessments of the implications of respacing and thinning for tree stability." Forestry: An International Journal of Forest Research 70.3 (1997): 233-252. Kyloe and Clocaenog described in: Hale, Sophie E., et al. "Wind loading of trees: influence of tree size and competition." European Journal of Forest Research 131.1 (2012): 203-217. Orange, Storrs and Torrington described in: Bunce, Amanda, et al. "Determinants of tree sway frequency in temperate deciduous forests of the Northeast United States." Agricultural and Forest Meteorology 266 (2019): 87-96. Other large tree motion data sets available at the time of writing (June 2021): https://doi.org/10.7910/DVN/FHJBYG http://www.hydroshare.org/resource/38ae9d9fb88d49f9ad2eed1ee07475c0 https://doi.org/10.5683/SP2/WZIKSR https://data.4tu.nl/articles/dataset/Tree_sway_of_19_Amazon_trees/12714989/1 https://doi.org/10.5285/657f420e-f956-4c33-b7d6-98c7a18aa07a https://doi.org/10.5285/533d87d3-48c1-4c6e-9f2f-fda273ab45bc

本仓库中的树木运动数据采自多个搭载不同传感器的观测站点。本数据集包含两类数据：(a) 来自27个站点的1小时采样数据（路径为/one_hour_samples/*.csv）。每个文件对应单棵树木，包含两列数据，分别代表树木运动的两个水平轴分量。此类数据未经过校准，且未附带时间戳（time-stamp）。文件名遵循格式：OwnersInitials_Site_treeID_resolution_SensorType（例如TJ_MY_7_4Hz_Strain），该文件名格式与TreeSummary.csv表格中的第一列完全匹配。

(b) 来自9个数据集的长期观测数据（路径为/site_name.zip/）。其中，风速数据与树木运动数据按每日单独存储，文件名标注了站点名称、树木ID或风速标识、年份、月份及日期，例如Kershope_T25_1990_5_23。每个文件的第一列为时间与日期，格式为YYYY-MM-DD HH:mm:ss.SSS；后续列依次为树木运动数据或风速数据，变量名中已标注对应单位。部分数据通过以已知拉力牵拉树木完成校准，此类数据的校准系数已在TreeSummary表格中给出。

澳大利亚数据集的风速数据存储于各对应树木的数据文件夹中，原因是各树木间距过大，无法共享一套风速测量装置。Orange、Storrs与Torrington站点的数据由倾角仪（inclinometer）采集获取。所有树木的单日数据均整合存储于单个文件中。

长期树木运动数据未经过预处理，因为用户的不同处理选择会对最终结果产生影响。需要特别注意的是，需从树木运动数据中移除由传感器漂移（sensor drift）引发的可变偏移量。详细说明可参阅随附的学术论文及其中引用的文献。

参考文献：
1. 阐述所有整合数据与必要预处理流程的学术论文：Jackson, T. 等 (2021) 《风力作用下的树木运动：数据综合研究》，Biogeosciences（《生物地球科学》）。
2. 澳大利亚数据集相关研究：James, K. 等 (2006) 《动态荷载下树木的力学稳定性》，American Journal of Botany（《美国植物学杂志》），doi:10.3732/ajb.93.10.1522。
3. Kershope与Rivox站点数据相关研究：Gardiner, B. A. 等，"Field and wind tunnel assessments of the implications of respacing and thinning for tree stability." Forestry: An International Journal of Forest Research 70.3 (1997): 233-252。
4. Kyloe与Clocaenog站点数据相关研究：Hale, Sophie E. 等，"Wind loading of trees: influence of tree size and competition." European Journal of Forest Research 131.1 (2012): 203-217。
5. Orange、Storrs与Torrington站点数据相关研究：Bunce, Amanda 等，"Determinants of tree sway frequency in temperate deciduous forests of the Northeast United States." Agricultural and Forest Meteorology 266 (2019): 87-96。

截至2021年6月发布的其他大型树木运动数据集：
https://doi.org/10.7910/DVN/FHJBYG
http://www.hydroshare.org/resource/38ae9d9fb88d49f9ad2eed1ee07475c0
https://doi.org/10.5683/SP2/WZIKSR
https://data.4tu.nl/articles/dataset/Tree_sway_of_19_Amazon_trees/12714989/1
https://doi.org/10.5285/657f420e-f956-4c33-b7d6-98c7a18aa07a
https://doi.org/10.5285/533d87d3-48c1-4c6e-9f2f-fda273ab45bc