Introduction
The purpose of this page is to give a handy reference to those who are interested in meteorological data in the area of the ING telescopes at the Roque de los Muchachos site. Data are shown in monthly graphs and maximum, minimum and mean value over the month are included for each meteorological parameter. These graphs are intended as an easy-read guide to some of the most interesting features. The data shown here for past years have been collected from the 1st of January until the 31st of December of each year. The current year is kept up to date on a monthly basis.

A brief description of the meteorological station
The meteorological station is divided into three groups of sensors: the main one is called ``JKT mast'' (as it is close to the Jacobus Kapteyn Telescope). Each of the other two telescopes has a smaller group of instruments, the William Herschel Telescope (WHT) and Isaac Newton Telescope (INT) local stations. Almost all the graphs refer to the JKT mast data, the only exception being the Local Humidity, which is reported for each group of sensors, as local differences in humidity make an interesting comparison.

The coordinates and heights of the three stations are as follows:

The following information is reported from the JKT Mast:

• Atmospheric Pressure, Wind Speed, Wind Direction, Wind Gust, Air Temperature, Soil Temperature, Borehole Temperature, Solar Radiation, Relative Humidity.

• Local Humidity.

• Local Humidity.

Local Wind Speed is not reliable due to the proximity of the sensor to the dome of the telescopes (the position of these anemometers was due to a financial deadline). Local Air Temperatures have been proved to always be consistently close to JKT mast values (within about 1 degree) so the reader can just refer to the JKT data which are included in this report.
For completeness, the characteristics of all the sensors have been included, even those whose data have not been reported.

Data storage
Each of the three stations has a Personal Computer which receives the data from the JKT Mast every minute and displays all the values; the data are stored on the PC disks every ten minutes.
Three new files (one for each stations' PC) are created every day at 00:00 UT and closed at 24:00 UT. These three files are transferred to a Unix database daily.
An example of the format used in the database is as follows:

This is a typical line of a file; the only spacing characters used are a colon (:) between hours, minutes and seconds, the blank space and the symbol #, this last being used when the data are not available; a header is added below to the same line to clarify the nature and units of each datum.

The JKT mast part of the file is common to the three sets of data, and therefore the JKT mast set has no LOCAL part (so # symbol appears in the space thus left free).
The three sets of data all contain the same JKT mast data which can differ from one to the other due to a small difference in the timing of data received, since the three PC clocks may have not been perfectly synchronized.

The graphs
All the parameters are represented in monthly graphs. In addition, the maximum, minimum and mean values over the month are given. All the values are shown as a continuous line on the graph (if no data are missing) except Wind Direction and Solar Radiation. For these two last parameters only discrete points or dashes are plotted, which makes the graphs easier to interpret.
Where the line of a graph is interrupted the corresponding data were missing or beyond the admitted limits for that parameter.
For each day and for all but the local parameters (i.e. not for local humidity graps), the most complete set of JKT mast data is used; in other words, the data for different days may come from different sets, from whatever telescope had less problems in recording the data.
In the Solar Radiation graphs the unit used for the maximum, minimum and mean values is W/m2, while the dashes of the graph represent the Integrated Solar Radiation over the day in J/m2. This is because the Integrated Solar Radiation would be expressed in
W x s / m2 = (J/s x s) / m2 = J/m2

Location of the sensors

On the JKT mast

• The main anemometer with wind speed and direction
• Hygrometer
• Thermometer
• Solarimeter
• Surface wetness sensor
• Dust monitor
• Barometer

• Soil temperature sensor
• Borehole temperature sensor

• Local anemometer with speed only
• Local external hygrometer
• Local internal hygrometer
• Local external temperature sensor
• Local internal temperature sensor
• Local surface wetness sensor
• Mirror temperature sensor
• Dewpoint sensor

• Solar panel
• Transmitter unit (only JKT mast)
• 12 V Electrolyte battery

• Personal computer and display screen
• Receiver unit
• Dewpoint meter controller

Anemometers

JKT main mast Anemometer

Local Anemometers

Vane

Hygrometers

Internal and external Hygrometer

Thermometers

Air Temperature thermometers

Soil and borehole thermometers

Solarimeter

Surface wetness sensor

Dust monitor

Barometer

Dewpoint meter

Assessment of data quality for the years 2000-2002
The whole system has proved to be quite good and most of the sensors have shown no evidence of malfunctions since installation. The sensors which have been tested by analysis of the data are: barometer, hygrometers, anemometers, thermometers. Visual checks have been performed on the dust monitor, wetness sensor, dewpoint sensor, solarimeter.

Pressure
Our sensor has been compared with a mercury barometer on site. Our readings are mostly low and the mean error (during 2000-2002) is estimated to be about 0.6 mbar, with a maximum of 2.2 mbar and minimum of 0.1 mbar. This error has been remarkably stable after 1998 (see graph).

Wind
Comparisons between the JKT anemometer and the one at the Meridian Circle, which has been working on site for over ten years, show good agreement, if one allows for small differences due to the different locations. The structure of our device leaves a gap from 356 to 359 (included) degrees. Any value falling within this interval is set to 0 degrees.
Some data were lost during the winter due to freezing of sensors mainly during January and December.
The wind speed by all the anemometers always looks quite consistent with handhold anemometer readings. A difference between the two of (say) 5 Km/h may be accepted as due to the handhold instrument always being lower than the fixed anemometers.

Temperature
Thermometers, either air or soil ones, always showed consistent values with each other. In November 2002 (19th) the borehole sensor failed and we have no data for the rest of November and December.

Solar radiation
A rough check of the Solarimeter readings compared with the Solar Constant always showed sensible values. The monthly graphs show a regular and smooth performance. The only error I could detect regards the moment the Sun appears after being obscured by cloud. The instrument cools when the Sun is hidden and gives slightly exaggerated values at the moment the Sun strikes it again. These jumps are intrinsic of the working principle of the sensor and have been filtered out by showing the Integrated Solar Radiation.
The shadow of one of the security ropes holding the mast falls onto the solarimeter for a few minutes around 08:30 UT. This, however, is not detectable on the graphs.

Humidity
Hygrometers, when checked during a stable period, proved to be quite reliable; their error, usually around 2%, gets bigger for either extremely low or high values (up to 15%). The WHT external hygrometer has been substituted on November 7th 2000 as it was reading about 3% less with very high humidity.

Power supplies
The solar panels work very well, the only problem is that the glass covers are quite fragile and during winter they are liable to get broken by ice falling from further up the mast or by water on the panel expanding when it freezes.
The batteries usually last well enough; a failure happened during a severe wind storm in December 2002, which blew off the solar panel from the INT and the local station ran out of power.

Data losses
The three telescopes' sets of data contain the same JKT mast data which can differ from one to the other due to a small difference in the timing of data received, since the three PC clocks may not have been perfectly synchronized.
The graphs of the JKT mast variables (i.e. all but INT and WHT local humidity graphs) were made using data from both set of files, WHT files and INT files, filling the gaps of one with data from the other. The resulting data loss during the year 2001 has been about 7.5%. During the year 2002 it has been about 8.7%. Most of the losses during 2002 were due to failure of the recording PCs and not to problems with the meteorological system itself.
The JKT set of data is kept only on the PC disk and is not available online.