7.1 Subprogramme AM: Meteorology

• 7.1.1 Introduction
• 7.1.2 Methods
• 7.1.2.1 Site requirements
• 7.1.2.2 Equipment
• 7.1.2.2.1 Instruments
• 7.1.3 Technical quality assurance
• 7.1.4 Data handling and quality control
• 7.1.5 Data reporting

7.1.1 Introduction

Since meteorological parameters are the most driving variables that effect ecosystems, their magnitude and changes in time should be well known to differentiate between anthropogenic perturbations and natural phenomena. In this context the need for phenological observations is evident (subprogramme phenology PH is under development) since meteorological data have to be calibrated by plant reactions on each site separately.

The objectives are:

• description of climatic conditions of IM sites and changes in these conditions
• detection of periods of extreme weather conditions and events that stress tree vitality (freezing of soils, late frost, drought, storm)
• preparation of a data base which fulfills the requirements of deterministic computer models that are capable of predicting ecosystem responses under future input scenarios.

Data from neighbouring monitoring stations meeting the set criteria (e.g. national meteorological networks) may be used for ICP IM purposes, provided that they can be shown to be representative for the IM site. The representativity of meteorological data with respect to various landscape types within the monitoring area must be carefully evaluated by the data originator. However, some site specific local data with relevance to hydrology, decomposition and soil classification are needed to interpret other measurements. Therefore at least soil and ground temperatures need to be measured at the IM site.

Traditional methods of data readings three times a day, carried out over decades, are useful for describing the climatic conditions of a given site. But for more sophisticated analysis of the weather dynamics the following variables should be measured on (quasi)continuously recording meteorological stations.

*For precipitation measurements some national standard heights exist. The resulting differences in precipitation amounts against the reference height should be tested and reported.

The height requirements are in accordance with WMO-Guidelines (1990) to ensure comparability with data from official weather stations and other monitoring sites.

Some compromise height specifications given for all radiation variables in brackets (..) should be considered as possible minima as far as the instrument is not shadowed by obstacles. At some sites radiation sensors have to be mounted near the top of the mast, which will complicate handling and maintenance of sensors.

7.1.2 Methods

7.1.2.1 Site requirements

Meteorological stations should be located in a clearing inside the IM site. Reflecting the large spatial variability especially of precipitation amounts and events the maximum distance to intensive monitoring plots should preferably not exceed 700-1000 m. According to WMO-Guidelines (1989) the minimum distance from the equipments to the next obstacle (tree) should be twice its hight, and the ground covered by short grass cut to lower than 10 cm.

Alternatively, towers that allow gradient measurements from above canopy to ground surface, are useful, but are also very expensive. Consequently, measurement towers normally are restricted to basic research stations. 

Data from neighbouring meteorological stations may be used for ICP IM purpose, provided that they can be shown to be also representative for the IM site. However, some site specific local data with relevance to hydrology, decomposition and soil classification are needed to interpret other measurements. Therefore at least soil and ground temperatures need to be measured at the IM site.

In this context the design of a field station is presented as an example. The station is running in remote forested areas, independently from mains. Its high reliability was successfully tested under various and rough climate conditions by the Bavarian State Institute of Forestry on its forest ecosystem monitoring network. Comparable instruments and sensors of any manufacturer may be used when fulfilling standard requirements.

7.1.2.2 Equipment

For mounting meteorological instruments at standardized heights, 10m masts are necessary (Annex X.1). Wind sensors are fixed at the top of the mast. Sensors at 2 m are clamped on arms (radiation to south, temperatures to north). Telescopic or some folding type masts facilitate installation, controlling and maintenance of instruments.

Data are recorded by means of a data logger, which is installed in a locker to prevent electronics from high humidity, preferably on or nearby the mast. The data logger should function reliably even at extremely low temperatures to minimize data losses. Note that storage on moving media (diskettes, tapes) is restricted to temperatures higher than -10°C, on memory cards (EPROMS) to temperatures below -20°C. The software programmes that are stored at EPROMS or down loaded from a notebook computer include, for each channel separately, sensor characteristics, reading and recording intervals, valid range and error setting, data conversion and compression as well as storage on memory cards. Procedures for on site checks of sensors and the whole system are carried out with separate programme cards or by temporarily connected portable computers. The power supply can be ensured by batteries connected to solar and/or wind generators. It should be noticed that the efficiency of a fully charged battery may drop down to about 50% at very low temperatures over longer periods. Stoppage can be counteracted by additional batteries recharged on mains.

The whole equipment should be protected against lightning and all the cables should be shielded against electromagnetic fields. Also the waterproofness of cable connections should be guaranteed by using high quality industrial standards. At some sites cables may be affected by mice.

7.1.2.2.1 Instruments

7.1.2.2.1.1 Precipitation

Besides the weekly or two-weekly precipitation amounts, which are measured in the framework of the precipitation chemistry subprogramme, precipitation intensity should be measured more frequently eg. to estimate evapotranspiration processes and to get information about the interception process.

While floating and tipping type buckets are restricted to liquid precipitation, the weighing type is appropriate for measuring all kinds of precipitation (snow, hail, mixtures of snow and rain) without heating.

The model presented in Annex X.2 is a Hellmann type gauge with the standard collection area of 200 cm², working on battery. Precipitation is measured by an electronic weighing compartment with a resolution of 0.01 mm and high accuracy. The linearized output signals are stored directly on memory cards.

It is of great advantage that loss of water through evaporation or through emptying the container is compensated internally. There is no need to correct zero shifting on computers.

mounting
The gauge is mounted on the upwind side of the mast with a horizontal distance of about 5-10 meters. The standard height is 1.3 m above ground which should be covered with short grass to prevent water from splashing in, and the collecting area is levelled plane.

maintenance
The need for maintenance is reduced to emptying the container in time and to system and battery check, which can be done during the routine check of the whole station (see below).

7.1.2.2.1.2 Temperature

The most common method to measure temperature is the use of platinum wire whose resistance changes with temperature. The widely used Pt100-resistor-thermometer with a basic value of 100 Ohms at 0°C is very appropriate for long-term and long-range monitoring of air and soil temperatures (Annex X.3). It is used to transfer IPTS 1968 (International Practical Temperature Scale) between instrument locations.

Accuracy should be ±0.3 K (WMO 1990, German Industry Norm DIN 43760 Class B ). When using Pt100 at different levels below and above ground, some elements should be calibrated simultaneously, and those exhibiting the most similar performance towards each other (zero point deviation, slope) should be applied facilitating corrections by software and especially heat flow calculations.

For air temperature the platinum coil is encapsulated in hard glass, for soil temperature it is built in a waterproof stainless steel tube.

mounting
Sensors for air temperature are installed in weather and radiation shields (Annex X.1). The cover and the lamellae of these shields should be cleaned regularly and checked for scratches to ensure appropriate reflection properties. Soil thermometers are to be placed in with good contact to the soil matrix.

maintenance
Since characteristics of the Pt100 are very stable (artificially aged before use) over long time, maintenance is normally restricted to renewal upon damage. Once a year zero-point calibration in ice water may be carried out for air temperatures. Soil sensors normally should not be removed.

7.1.2.2.1.3 Relative humidity

In battery supplied field stations two main types of passive humidity sensors with low power consumption are widely used, often combined with PT100 temperature sensors, namely: a) hair hygrometers and b) capacitive sensors. Accuracy should be within +/- 3%.

mounting
Like thermometers humidity sensors are to be mounted inside a weather and radiation shield, prevented from precipitation, splashing water and direct solar radiation.

a) hair hygrometers (Annex X.3)
The measuring element consists of a number of treated natural hairs or synthetic fibres that change in length when the relative humidity changes. This change is transmitted to a potentiometer creating an electrical output to a recorder/display. Natural treated hair elements are appropriate for most nemoral and boreal forest ecosystems, covering large temperature (-35°C-+70°C) and relative humidity ranges (10-100 % rel. humidity).

maintenance
Hair elements dry out over longer periods with low air humidity resulting in too high values. This degeneration can be reversed by exposing the elements to warm saturated air. This can easily be done at the station by wrapping the case with a wetted cloth for about one hour. The humidity value then should be 97 %, otherwise, it has to be adjusted by the setting screw.

In forest ecosystems hair elements can be periodically exposed to pollen causing significant errors. If the weather shield is insufficiently protecting the hair, a protection device may be used which has to be checked and cleaned regularly.

calibration
The above-mentioned procedure may also be used for a one point calibration twice a year. Comparison with a portable standard aspiration hygrometer is useful to examine the lower part of the measuring range (40-50 % rel. humidity) and the long-term stability. New hair elements have to be calibrated in a humidity chamber. In foggy air it is self-calibrating.

b) capacitive sensor (Annex X.3)
The capacity of a polymer film changes when absorbing water vapour. These capacity changes are detected by electrodes and converted to electric signals. Since they cover a temperature range from

-20°C to +80°C and a humidity range from 0-100 %, they might be more appropriate than hair hygrometers in warmer and dryer climates.

maintenance
Stability of the endpoint can be checked by exposing the sensor to warm saturated air as described above. The response time should be noted and compared with the users manual to verify the permeability of the particle filter. Adherent dust affects the response time and the filter has to be changed. Since the sensor housing can be sealed by ice layers during long cold periods, another sensor type may be selected under such winter conditions.

calibration
Twice a year a two-point calibration procedure should be carried out using the manufacturer's calibration set with saturated salt solutions.

7.1.2.2.1.4 Wind speed and direction

With a wind transmitter, both horizontal wind speed and direction can be measured by anemometer and wind vane. Commonly propeller and cup anemometers are used whose angular velocity is directly proportional to wind speed and is transmitted by signal generators of different types.

For example (Annex X.4), the ball bearings of the cup anemometer are coupled to a slotted drum, which is scanned opto-electronically which is insensitive to electromagnetic fields. The pulse frequency is proportional to wind speed. Quality criteria are a large measuring range (<1-50 m/s), a low starting speed (<0.3 m/s) and a distance constant of about 2-5 meters.

The ball-bearings of the wind vane are coupled to a code disk, from which the current code is detected opto-electronically. Since the response of a wind vane to sudden change in wind direction is characterized by overshooting and oscillation about its true position, the damping ratio should be in the range of 0.3-1.0. Satisfactory resolution and linearity in wind direction are 2°-5°.

mounting
Through the first few (tens) of meters above the ground the wind speed varies considerably due to friction. For this reason wind transmitter is to be mounted at the top of the mast at 10 m (at standard height). The case marking and the handle of the wind vane are aligned to compass north.

Where the standard exposure is impossible, the wind speed at 10 m may be derived from other heights by using the formula of Hellmann according to WMO-Guidelines.

maintenance
The alignment of the wind transmitter should be checked regularly using a theodolite at a permanently marked fix point.

Unless heated, ice and snow can accumulate on cups leading to higher starting torque or cessation. This can be accepted in wintertime when the evapotranspiration is very small. Slots that can be clogged by dust should be regularly cleaned.

Depending on the instructions of the manufacturer, wind sensors are checked after several years of exposure in a laboratory on signs of wear of the ball bearings as well as after meteorological events (cup deformation by hail).

7.1.2.2.1.5 Global radiation

Direct solar radiation and diffuse sky radiation to a horizontal plane surface comprising the spectral range from 0.3 to 3.0 µm are referred to as global radiation. It is measured by pyranometers. Two measuring principles commonly used are:

a) black painted disks that absorb incoming short wave radiant energy, generate a heat flow through a thermal resistance to the pyranometer body. The temperature difference between disk and pyranometer body is converted into voltage.

b) differences in temperature between absorbing black and reflecting white areas are transformed with thermoelements into a proportional voltage.

Both sensors are shielded by glass domes that protect against wind, rain and energy losses but allow transmittance of the incoming short wave radiation. Where the receiver is not completely sealed, it has to be protected by a desiccator against condensed moisture.

In this context a pyranometer of the a) type is recommended (technical details in Annex X.5). It fulfills the WMO requirements (ISO 9060) for a secondary standard and is calibrated by direct or indirect comparison with a primary standard. It defines geometry and temperature response characteristics as well as spectral sensitivity, stability and linearity over time.

mounting
Above the plane of the sensor the location should be as free as possible of any obstructions, which may shadow it at any time of the year. The elevation of any obstruction should not exceed 5° over the azimuth range between earliest sunrise and latest sunset to achieve correct measurements of direct solar radiation while diffuse radiation is much less affected.

Use of masts that allow frequent control and maintenance, causes conflict with the above-mentioned requirements, and should be minimized. In practice, the arm on which the sensor is mounted, should be oriented to the south (the mast itself to the north pole). As far as polar diagrams of directional radiation responses are available, the low error region of the sensor may be turned to the equator. The accurate horizontal levelling of the thermopile surface is done by levelling screws and a spirit level.

maintenance
Pyranometers should be inspected daily as recommended by WMO. In large forested sites intervals of one week should be sufficient at least outside periods of flowering and snowfall.

During inspections the glass dome of the instrument should be wiped clear and dry as gently as possible not to alter transmission characteristics. Frozen deposits are removed using a de-icing spray and the glass dome is cleaned (sa). The desiccator material (usually silica gel) should be renewed when discolouring.

calibration
Apart from calibration procedures by standard pyrheliometers and pyranometers in qualified radiation centres, routine checks of calibration factors using the sun as radiation source, should be carried out at least once a year. A reference standard or a travelling working standard, preferably of the same type of instrument, is mounted side by side at the mast, allowing simultaneous recording over one or two days. The means over several time periods of both sensors are used to calculate a calibration factor of the sensor to be checked. Alternatively the pyranometer may be exchanged by a similar calibrated one while recalibrating indoors.

7.1.2.2.1.6 UV-B radiation

The possible damaging effects of UV-B radiation, caused by thinning of the stratospheric ozone content, to biological systems has been discussed. Because the slope of energy flux in the wavelength band of 280-315 nm is very steep, accurate values can only be measured by scanning the spectrum with high resolution (1 nm). The necessary instrumentation is, however, costly and not appropriate for field stations.

Cheaper types of instruments are the broadband UV-B radiometers which cover a bandwidth of 20 to 40 nm or the whole UV-spectrum with varying peak wavelength and varying spectral response, depending on filter characteristics.

A solution considering costs, practicability and scientific value may be the use of narrowband sensors with defined peak wavelength and bandwidth. In Annex X.5 such a sensor is presented, the peak wavelength adjusted to 306 nm which is derived from human skin and solar spectra. To get a high accuracy the sensor is heated to 40°C. Consequently there is high power demand that hardly can be satisfied at remote field stations.

Mounting and maintenance procedures are done according to 7.1.2.2.1.5. Calibration is based on standard lamps by special laboratories.

Since UV-B radiation energy at ground level is mostly determined by stratospheric processes, data from official weather or environmental network stations some tens of kilometres away can be used, taking into account the altitudinal dependency of radiation received.

Using non-scanning type sensors, peak wavelength and bandwidth have to by standardized to allow transboundary comparisons.

7.1.2.2.1.7 Photosynthetic active radiation

The radiation spectrum of 400-700 nm that is used by plants for photosynthesis is referred to as photosynthetic active radiation, it amounts to about 50% of global radiation. So called quantum sensors (Annex X.6) count the number of photons in this spectrum falling per unit time and area through a spectral filter on a blue enhanced photocell, regardless of their energy. This is called the photon flux density in units of µEinstein or µmol photons(quanta)/m2/s.

The sensor should have a working range from 0 to 5mmol photons/m2/s. Quality criteria are high linearity (1%), long term stability (<±2%) and small temperature dependence (±0.15%).

Mounting is done according 7.1.2.2.1.5 and maintenance is reduced to cleaning the sensor surface and to checking the levelling. Although routine calibration should not be necessary the ageing of filter material and photocells under specific field conditions should be taken into account.

7.1.3 Technical quality assurance

When designing an automatic weather station the site specific weather conditions need to be kept in mind. The sensors and instruments need to be reliably running and give the reported accuracy even under extreme weather conditions. The performance of the instruments and sensors should be certificated, and the instruments should have clear instructions about calibration procedures and recalibration intervals both at field and laboratory. The lifetime of components should be known so that the exchange of spares can be done in time. A modular construction of instruments enables easy replacement of spares and reduces data losses.

Once or twice a year an integral check of electronic components such as cables, connectors and analog/digital converters should be carried out by simulating electronically the output signals of sensors according to their specifications. For example, given 100 and 88.22 Ohm to a Pt100 signal cable, 0°C and -30°C resp. should be seen on the screen. This procedure ensures that the sensor output is correctly transmitted to data logger input and converted to meteorological values.

At least weekly field checks should be carried out by well trained personnel, examining power supply and the correct operating of data logger and sensors. The specific maintenance needs are reported below. A formalized log-book, containing all details to be checked, facilitates the maintenance and the evaluation of data, too.

7.1.4 Data handling and quality control

measuring and recording intervals
As pointed out earlier the weather station should work (quasi-)continuously to enable more in depth analysis of for example plant-weather-relationships. This means, that measuring should be carried out at time steps of 1min. To detect also extremes of very dynamic variables like wind, where the speed of damaging gusts can reach a multiple of even a short time average, time resolution should be one second. For that reason interval measurements in time steps of one hour for example are not sufficient regarding wind speed.

Data recording intervals may not exceed 60 minutes comprising averages and minimax of variables with the exception of wind direction, where frequency distributions are needed.

data quality management
Down loading of records or exchange of memory cards is related to the capacity of storage media, but should occur at regular intervals ( 4 weeks), preferably once a week during station service, to identify errors not detected by field workers. Therefore data evaluation should also be carried out regularly (consequently at least monthly).

After transmission, data sets should be routinely examined on time consistency, missing values and error settings and edited if necessary. A subsequent plausibility routine should check on:

• working range exceedance of each sensor and sensor specific error data (e.g. zero offset of pyranometers at night time by temperature gradients)    
• internal consistency errors (for example: minimum temperature exceeds average)    
• exceedance of absolute limits (e.g. wind direction), probable limits (absolute minima and maxima) and average frequency distributions which can be derived from long term records of the next official weather station, taken altitudinal gradients and specific orographic conditions into account.

Data that are formally invalid are to be flagged for particular inspection.

For the subsequent scientific analysis construction of diagrams on display is a very useful means: time sequences of each variable let detect erratic changes of normally conservative variables (temperature, humidity) or stability of normally dynamic variables (wind direction).

Comparisons with related variables facilitate the decision, if there are technical problems or extraordinary weather events. Examples: the windvane may be fixed by icing while the anemometer is running; if there is a rapid decrease of air temperatures (20K) in wintertime during few hours then a change in predominant wind direction should have occurred.

All evident errors, missing and doubtful values, that are excluded from further calculations, as well as corrected (interpolated) values are flagged on the database.

In the sense of a final examination, preliminary calculations of averages, minimax and distributions on daily base should be carried out, allowing comparisons with corresponding data of neighbouring official weather stations. Significant non-explainable differences related to distribution, course or range of different variables, should result in rechecking the data set. Principally the feedback time between field workers and scientist and vice versa should be as short as possible.

7.1.5 Data reporting

Meteorological variables are reported on monthly bases. Statistical status are mainly sums (precipitation), arithmetic averages as well as absolute and averaged extremes, that are flagged according to the table below.

Predominant wind direction is reported as mode. Individual recordings are classed for example to 12 sectors of the windrose, each covering 30 compass point equivalents in degree, starting from NNE=30° (15°-45°) to N=360° (345°-15°). Calm (wind speed<0.2m/s) is reported as zero.

Even if only monthly values are to be reported to the IM programme centre, the original data are to be stored by the data originator and to be made available on request. Files containing validated hourly or at least daily values, should be deliverable upon request for sophisticated analysis.

mandatory information

optional information

status information

Example files

AM example Excel file
AM example ASCII file

• File identifier SUBPROG states the subprogramme.
• MEDIUM is given for temperature values as AIR or SOIL, for other parameters it is left blank.
• LEVEL is given as the absolute height/depth of the measuring equipment from the ground (cm).
• Spatial pool SPOOL refers to the number of individual recording devices used for each parameter.
• For each parameter several values are reported, averages, average maximums etc., the corresponding status flags needs to be included.
• Sampling year and month are given as YYYYMM , day field is left blank.

Additionally the beginning and the length of ecologically important periods, that cannot be calculated from monthly data, should be reported. (Reporting of these values is done separately in free format).

For the beginning of the vegetation period the date is reported where mean air temperature exceeds the threshold value of 5°C for at least 5 consecutive days. The length of this period is then calculated by counting the number of days to that date on which mean temperature remain under 5°C.