Strain data measured on the trunks of 19 trees in the Danum Valley, Malaysia. It also contains wind speed data measured using anemometers attached to emergent trees nearby. This data can be used to extract the fundamental swaying frequencies of these trees, and to quantify the relationship between wind speed and bending strain and so estimate wind damage risk. This data would be useful for anyone studying the interaction between wind and trees. The tallest tree in this data set is approximately 51 m tall. The nearby field centre is located at (4°57'53.0"N 117°48'18.3"E) and the measured trees are part of the 50ha Smithsonian plot and the 1ha intensive carbon monitoring plot 1. 

Strain (extension / original length) was measured using pairs of strain gauges attached approximately perpendicularly to each other in order to capture the two axes of motion. Strain was monitored at 4Hz and all times are given in UTC. Wind speed was monitored at different resolutions according to the instrument. The anemometers were mounted on metal poles attached to the trunk and boomed out approximately 1.5m. 

The file 'Trees.csv' gives tree species ID and a census ID for each tree which identifies it in the Smithsonian census. 
The file 'StrainGauges.csv' gives the arm length, calibration coefficient and mounting angles necessary to process the raw data.
Most users should not need to process the data, since the raw data (suitable for spectral analysis) 
and the processed data (suitable for time-series analysis) are both deposited.

For the matlab scripts used to analyse the data visit (github.com/TobyDJackson/WindAndTrees_Danum)
Please email me if you have any questions: tobydjackson@gmail.com

The strain data is organised by data logger (C1 - C5):
C1 - CR1000 with 3 pairs of SGs (arranged in pairs on the same tree at 1.3, 25 and 35 m heights, with the top pair being attached to a 1st order branch) 
C2 - CR1000 with 4 pairs of SGs on four trees
C3 - CR1000 with the first 3 pairs of SGs on one tree (at 1.3, 19 and 30 m) and 1 pair on another. 
C4 - CR23x with 6 pairs of SGs on separate trees and logged using a trigger system. 
C5 - CR23x with 6 pairs of SGs on separate trees and logged using a trigger system. 

The wind data is:
C1_cup - 0.1 Hz resolution wind speed data measured at 25 and 35 m heights on tree 1, using cup anemometers (Vector Instruments A100LK/5M)
C3_cup - 0.1 Hz resolution wind speed data measured at 19 m on tree 6, using a single cup anemometer (Vector Instruments A100LK/5M)
C3_sonic - 1 Hz resolution wind and direction data measured at 30 m height on tree 6, using a 2D sonic anemometer (Gill Sonic 1)

Overall, the data quality from the CR1000s is good, but that from the CR23xs is poor. Reasons for this are
discussed in the associated publication. The wind data requires no processing.

The raw data go through a number of calibration and processing stages:
1. Raw data - measurements in mV directly from the data logger *****
2. Calibrated data 
3. Filtered data - long term drift is filtered out using a Butterworth band pass filter.
4. NE data - strain on the North and East facing sides of the trunk 
5. MaxStrain data - a combination of the two channels to give the maximum strain at each time step.

***** This is the version available online. Other versions are available on request.