### Data use _Until further notice the data at this repository remain publicly available and can be used for non-commercial purposes on the condition that this repository is cited in the list of references of publications as the source of the data, including the DOI to ensure readers can access them._ For other uses or collaborations requiring supply of customized versions of the data files or help with the interpretation or use of the data, please contact [Pedro J. Aphalo](mailto:pedro.aphalo@helsinki.fi). ### Location and equipment This dataset is being acquired at the Viikki campus of the University of Helsinki, Helsinki, Finland. (lon = 25.01673 E, lat = 60.2253 N) The weather station, instrumentation and variables are currently described in detail at [a separate website][1]. A summary is given below. A Campbell Scientific CR6 datalogger with a CDM-A116 analogue imput module is used to acquire and log the data. The datalogger has both digital and analogue inputs. The anologue to digital conversion (ADC) is done with 24 bits resolution and autoranging is available. Data is acquired from most sensors once every 5 s and summaries computed and saved at 1 min-, 1 h-, and 1 d intervals. Soil sensors are read less frequently, and summaries saved at 1 h intervals. In the daily data, in addition to means, standard errors, maxima and minima are logged for most variables, and histograms for radiation data are computed in the logger. In daily data, times for maxima and minima are recorded. Each of the sets of summaries, 1 min, 1 h and 1 d, are stored in different *tables* (storage space) in the logger's CRBASIC program and downloaded as separate text files. A fourth table is used to store radiation data at the rate they are acquired. Data are downloaded on site through a USB connection to a laptop computer. _**Table** Sensors in the weather station. *signal* is how the datalogger receives the information. Sensors with mV (millivolt) output lack built-in amplifier and provide a raw electrical signal. Sensors with V (Volt) output have a built-in analogue amplifier. Signal in ohms ($\Omega$) is the sensor's electrical resistance. SDI-12 is a digital serial communication protocol. In the case of analogue sensors, the digital conversion is done by the datalogger. For digital signals the conversion is done in the sensor. Most SDI-12 "sensors" like the WXT520 "weather sensor" contain sensors for multiple variables. SDI-12 communication allows the data from difference sensors to be sent sequentially using the same wires and logger inputs. *since-to* gives the period when sensors have been operative. *time step* refers to data acquisition. *logging* refers to the time step for data storage: *rt* = real time or same as data acquisition; *min* means storage of the average of acquired data for each minute, and possibly also maximun and minimum; *h* as *min* but once per hour. When data are logged once per minute, hourly averages are in some cases computed from the 1-min averages. For radiation data, daily histograms are also logged._ | Sensor type | make | variable | qty. | signal | since-to | time step | logging | |:-----------------|:------------|:--------------------------|:-----|:-------|:--------------|:----------|:--------| | UV-Cosine (UVB) | sglux | UVB irradiance | 1 | 0-5V | 2020/05 | 5 s | rt, min, h | | UV-Cosine (UVA) | sglux | UVA irradiance | 1 | 0-5V | 2020/05 | 5 s | rt, min, h | | UV-Cosine (blue) | sglux | Blue-Violet irrad. | 1 | 0-5V | 2020/05 | 5 s | rt, min, h | | SKR-110 R/FR | Skye | red irradiance | 1 | mV | 2017/06 | 5 s | rt, min, h | | SKR-110 R/FR | Skye | far-red irradiance | 1 | mV | 2017/06 | 5 s | rt, min, h | | LI-190 quantum | LI-COR | PAR (total PPFD) | 1 | mV | 2016/01-2022/06 | 5 s | min, h | | CS310 | CampbellSci | PAR (total PPFD) | 1 | mV | 2021/06 | 5 s | rt, min, h | | SMP3 pyramometer | Kipp | global radiation | 1 | 0-1V | 2016/01 | 5 s | min, h | | BF5 | Delta-T | PAR (total PPFD) | 1 | 0-2.5V | 2017/06 | 5 s | rt, min, h | | BF5 | Delta-T | PAR (diffuse PPFD) | 1 | 0-2.5V | 2017/06 | 5 s | rt, min, h | | WXT520 | Vaisala | Air temperature | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Air humidity | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Wind speed | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Wind direction | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Atmospheric pressure | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Precipitation, rain | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT520 | Vaisala | Precipitation, hail | 1 | SDI-12 | 2016/08-2021/06 | 5 s | min, h | | WXT530 | Vaisala | Air temperature | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Air humidity | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Wind speed | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Wind direction | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Atmospheric pressure | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Precipitation, rain | 1 | SDI-12 | 2021/06 | 10 s | min, h | | WXT530 | Vaisala | Precipitation, hail | 1 | SDI-12 | 2021/06 | 10 s | min, h | | SoilVUE10 | CampbellSci | Soil moisture profile | 3 | SDI-12 | 2020/05 | 12 min | h | | SoilVUE10 | CampbellSci | Soil temperature profile | 3 | SDI-12 | 2020/05 | 12 min | h | | SoilVUE10 | CampbellSci | Soil elect. cond. profile | 3 | SDI-12 | 2020/05 | 12 min | h | | SoilVUE10 | CampbellSci | Soil permittivity profile | 3 | SDI-12 | 2020/05 | 12 min | h | | CS655 | CampbellSci | Soil moisture | 3 | SDI-12 | | 12 min | h | | CS655 | CampbellSci | Soil temperature | 3 | SDI-12 | | 12 min | h | | CS655 | CampbellSci | Soil elect. cond. | 3 | SDI-12 | | 12 min | h | | CS655 | CampbellSci | Soil permittivity | 3 | SDI-12 | | 12 min | h | | 107 | CampbellSci | Soil temperature profile | 3 | ohms | 2020/08 | 5 s | min, h | | CSmicro LT02 | Optris | Surface temperature | 2 | 0-5V | 2020/11 | 5 s | min, h | The preprint ["Anticipatory responses to drought by plants: What are the environmental cues?"][2] describes some of the methods and makes use of some of the data from this data set. Over the years new sensors have been added and some have been retired. Which quantities have been logged also varied depending on changes to the program running in the logger. All data have been reprocessed on 2023-11-04, and new versions of all files uploaded. The data are processed in R, and checked for outliers and other problems. The scripts used have been updated, but all data files here have been obtained with the same version of the scripts. Radiation sensors have been calibrated on site against a spectrometer a few times each summer. _The order of the columns can differ between versions of the data, but variable names have remained in most cases unchanged._ All 1 min interval data files contain the same 57 columns, possibly filled with NAs at some ranges of dates. All 1 h interval files contain the same 235 columns, possibly filled with NAs at some ranges of dates. Data logged at 1 min intervals are available in one file per year as R's `.Rda` and as `.csv` files. The CSV format text files are gzip-compressed, use a dot as decimal marker, a comma as separator, newline as end of line and are encoded in UTF-8. In CSV files times are in UTC (not EET, the local time coordinates). A companion file `.met`, is a text file with metadata. The `.csv.gz` files need to be uncompressed before use. Most Linux distributions include the `gzip` (GNU ZIP) command line program. The program `7zip` (https://7-zip.org/) has a graphical user interface and in available for several Linux distributions, OS X and Windows. In some cases, the decompression takes places automatically. The data sets include in addition to the measured data, values for the position of the sun computed with R package ['photobiology'](https://docs.r4photobiology.info/photobiology). ### Warnings and caveats _The data set has some gaps, and the standards of calibration and frequency of checking are lower than for weather stations managed according to WMO standards. The special significance of the data is in its high frequency (5 s/1 min) and the multiple wavebands of sunlight being measured simultaneously._ **WARNING 1 Winter time radiation data** These data are suspect, due to lack of air blowers in radiation sensors and consequent possible accumulation of snow. Since October 2022 the BF5 sensor has heating enabled and will be less affected by snow and ice than other sensors. Because of its shape, the newer CS PAR sensor is less prone to this problem than the LI-190 sensor. On some years, data was not acquired during winter time. A battery failure caused another shorter interruption. **WARNING 2 Disturbances to radiation data from birds** The radiation data has occasionally very brief periods, mostly individual time points in 1 min data, of _darker_ readings caused by birds sitting on top of the sensors. In most cases it should be possible to detect these by comparison to other radiation sensors. **WARNING 3 Disturbances to soil profile sensors in the Summer of 2023:** Measurements were disturbed repeatedly in spite of requests to be careful and warnings that measurements were being disturbed. Those in charge of a neighbouring experiment (under a P.I. from the Centre of Excellence in Tree Biology, University of Helsinki) repeatedly watered the area where the weather station and the soil-profile sensors are installed. Subsequently in the morning of 7 July a tractor carrying out tasks related to this same neighbouring experiment drove into the weather station plot and directly over one soil profile sensor. **Thus, the soil-profile data for June, July, August and September are suspect and have been set to `NA` or not available. I may restore some of these data in future versions.** ### News [2022-01-03] Added data for UVA1 and UVA2 irradiances. Slightly adjusted the calibrations for red, far-red, blue, UVA and UVB and PAR irradiances for years 2020 and 2021, using additional spectral data. From 2021 onwards PAR data in column `PAR_umol_CS` should be used in preference to `PAR_umol_LI` as the 30-years' old LI-COR quantum sensor data is unreliable. [2022-01-15] Added data for rain (mm per minute) in column `rain_mm_min`. Only available starting from **2021-06-09 14:00:00**. [2023-01-16] **1 min interval files:** added files for year 2022. Small errors with times and calibrations were retrospectively corrected for some of the earlier years, and new versions of all files uploaded. In 2022 we had a battery failure and firmware bug problems, that caused some gaps in the time series, shown as NA (not available). [2023-01-16] **Hourly data files** added, computed from the 1 min data for most variables, except for soil water profile and soil temperature profile data that are logged once per hour. [2023-06-13] Uploaded data up to 2023-06-12. Renamed the file containing hourly data by replacing the current year by "latest", so that the file name remains unchanged in the future. [2023-11-04] Variables `PAR_umol` and `PAR_umol_sd` containing data acquired with an old LI-COR LI-190 quantum sensor have been renamed `PAR_umol_LI` and `PAR_umol_LI_sd`. Starting from the data files released today, variables `PAR_umol` and `PAR_umol_sd` contain the best data available, from different sensors, depending on the date. The soil profile data acquired between 2023-06-01 and 2023-10-01 have been marked as not available (`NA`), because the ground where sensors are located was accidentally watered several times in June and early July. Heavy rains in October saturated the soil profile, and subsequent measurements are expected to be once again valid. An overcorrection in the zero calibration of the PAR sensor from Campbell Scientific (CS-310) by -50 umol m-2 s-1 was fixed and the data recalculated for 2022 and 2023. The files uploaded replaced all previously uploaded ones. [2024-04-12] Logger OS and CDM-A116 OS updated to most recent versions. Logger program updated adding a new data table that samples radiation sensors once every 5 s, thus saving all radiation data as acquired to a separate file. This table is active only if PAR > 1 umol m-2 s-1. Data from the pyranometer are not included in this table, as this sensor is slower (response time 12 s). [coming soon] UV-A, UV-A1 and UV-A2 data corrected for the temperature coefficient of the sensors. New sensors will be delivered soon for more accurate measurement of UV-A1 irradiance, and for the measurement of green light irradiance. I intend to increase the measuring frequency further, to at least 1 Hz. [1]: https://viikki-stn.r4photobiology.info/ [2]: https://doi.org/10.32942/osf.io/ypqea