**Location of Sampling Stations**: Flumes are located at the base of each catchment ASD02/N04D; ASD04/N20B; ASD05/N01B; ASD06/N02B.

**Frequency of Sampling**: Stream gage height is recorded every five minutes on the CR-21X datalogger (Campbell Scientific Co.) up to 2012 at which point the depth logger was switched to a SUTRON Accubar Constant Flow Bubbler with integrated data logger. Each record also includes Julian day and time (discontinued in 1999 recorded hourly (see data set AWT02). Data were initially dumped at approximately one to two week intervals from the CR-21X memory to cassette tape (up to about 2000) or laptop computer (2012). Currently data are relayed via wireless to Bushnell Hall every 6 hours and are backed up regularly. A computer program is used to reduce and summarize the five minute values into daily summary values and stormflow summary statistics.

**Methods**:

Until 2013, gage height is sensed by pressure transducers (Druck Model PDCR 10/D) and recorded on the CR-21X datalogger (Campbell Scientific Company), or by a SUTRON Accubar Constant Flow Bubbler with integrated data logger. In 2013, pressure transducers were replaced with air bubblers with an integrated pressure sensor (YSI WaterLOG Bubbler/Pressure Sensor H-3553) and data are recorded on a CR800 datalogger (Campbell Scientific Co.). Conversion to stream discharge requires two steps:

1) Correction of measured gage height to actual gage height using direct measurements of stage made approximately three times per week with a ruler at a reference point at each flume. Until 2018, corrections were applied annually by conducting a linear regression between time-matched manual (x) and bubbler (y) measurements of gage height using all data collected in a year. All bubbler measurements of gage height from that year were then corrected with the resulting regression equation. Starting in 2018, a two-point regression method has been employed by examining the linear relationship between two consecutive points with manual and bubbler measurements of gage height. Manual measurements were collected approximately three times per week, therefore corrections were applied with the same frequency. All these corrections have been applied separately to each sampling station. Each sampling station is tested for high flows one time per year by plugging the stilling well and filling the well with water to artificially increase gage height. Water level in the stilling well is measured with the aid of a clear plastic tube running through the plug that fills to the same height as the well.

2) Translating gage height to stream discharge using a rating curve.

Two rating curves employ geometric relationships which are assumed to be valid at certain stage heights. The relationship used for calculating discharge at gage heights > 18.25 cm is:

Qm3/s = 4.64 x 10-5 * s^2.587

The relationship used for calculating discharge at gage heights between 0-18.25 cm is: Qm3/s = 6.49 x 10-5 * s^2.4714

Where Q is discharge in cubic meters per second and s is gage height in cm. These equations were derived using procedures in Replogle, J. A., H. Reikerk, and B. F. Swindel. 1978. Water monitoring in coastal forest watershed studies. IMPAC Report 2, Vol. 3, No. 2. Southwestern Forest Expt. Station, USDA, Gainesville, Florida.

**Form of Data Output**:

Data are downloaded every one to two weeks until 2012 and has been downloaded at least every 24 hours since then.

Once corrected, data is stored on the LTER network. Daily and storm flow (five minute value) summaries are available. Before 2005, only data for storm events were saved. Due to decreases in limitations of computer space, we are not saving complete hydrologic records for each watershed, and allowing individual users to convert height to storm peaks themselves for specific events they are interested in. Gage height data can be converted to discharge by using the equations above.

**Summary of All Changes**:

1. Change all the blanks cells for missing values before 2013(on Jan & Feb, 2021), after 2013( on Aug, 2018)

2. Set the time min max to missing if there is zero flow during a day for all year’s data (on Jan & Feb, 2021).

3. For data quality flags, on the flag columns, before 2013, put a missing code. After 2013 change blank to ‘zero’, what those meaning are: 1 = flag, 0=no flag, ‘.’ = missing value (Aug, 2018, Jan Feb, 2021)

4. For variables of volume, height, Sheight, CorrectedSheight, Discharge, RecHour etc, if any of the value is null or blank, willl set them for missing value '.' (Feb 2021)