V. Description of early warning procedures in the German gamma dose rate network

From Intamap

V. Description of early warning procedures in the German gamma dose rate network

The measuring sites are queried daily by modem during routine operation and their measured values are examineed. This routine is automated to a large extent. Today intrinsic mechanisms of the measuring device perform automated procedures for the on-site validation of the data. As soon as a dose rate threshold, which is computed for each site individually, is exceeded, a message is sent to one of the six central servers. In an emergency situation the acquisition frequency can be increased up to 10 minute intervals allowing to follow the atmospheric transport and deposition of radioactivity almost on-line.

Step 1: Routine data checks

In the German gamma dose rate network under normal conditions, data from all monitoring stations are queried from the central server unit once every day during the night. Data from each individual monitoring station are checked using automatic algorithms which are optimized to identify conspicuous data. Those data will be marked as “to be validated" and have to be interpreted manually by experienced personnel.

The time series of data from conspicuous monitoring stations have to be checked considering available information e.g. from neighbouring dose rate stations, from other radiological networks, meteorological stations or from precipitation radar system. As result of this manual procedures all data marked “to be validated" have to be classified as follows. If the result of the validation procedure is that the data are valid, the status will be changed to “plausible", otherwise it will be changed to “not plausible". The validation procedure lasts as long any data with status “to be validated" are available on the central server. This classification procedure is mainly based on expert knowledge from the time evolution of measured data in normal situation (e.g. rain events: wash-out of natural air activity of radon daughters). Detailed knowledge of typical technical perturbations of the measuring devices is needed as well.

Step 2: Spontaneous transfer of increased dose rate data

The GDR network is used to detect increased levels of radioactivity. In figure 12 two effects are shown together to explain the difference between increased levels of radioactivity due to the wash-out of naturally occuring radon daughter products and data obtained after the Chernobyl accident.

As can be seen from figure 12, the mean level of radioactivity is about 80 nGy/h at the station Karlsdorf compared to 100 nGy/h at the Schauinsland mountain. The maximum values due to wash-out of radon daughter products and after the Chernobyl accident are close to 150 nGy/h. In both cases the ground level is more than doubled. The difference between both examples is the different decay time. Radon daughter products decay with a half-life of about 2 hours (which is the characteristic time taken for the activity of a particular radioactive substance to decay to half of its original value). In the case of the Chernobyl accident the half-life of Cs-137 takes 30 years.

To detect any increased level of radioactivity, in all GDR monitoring stations mechanisms are implemented to identify data which exceed a given alert threshold. These data are immediately transmitted to the central server unit of the monitoring network, generating a so called spontaneous report. The alert threshold for a given station is defined individually by typical mean values of the dose rate at the station and the variability of measured data in normal situations. Thus, the threshold depends on the sensitivity of measurement system as well as site specific influences.

The threshold values of the Geiger-Müller (GM) tubes used in the BfS gamma dose rate network are in the range between 0.05 an 0.1 µGy/h above the mean dose rate in normal situations.

Step 3: Alert function of GDR network

The software of the central server unit of the monitoring network performs every 10 minutes an automatic early warning check. An network internal alert is automatically generated, whenever spontaneous reports form neighbouring stations have been reported in a certain time interval.

For the German monitoring system, the mean distance of neighbouring monitoring stations is about 15 km. Therefore, a so called space-time window was defined as follows: the distance between stations with enhanced data is less then 30 km and the spontaneous reports from the stations are registered within 1 h.

Step 4: Evaluation of the situation - Generation of an alarm

If a network internal alert is generated, the staff member on duty has to evaluate the situation as fast possible. For this purpose, the time evolution of the data of the monitoring stations with enhanced data has to be checked. Again additional data has to be considered e.g. from neighbouring dose rate stations, from other radiological networks, meteorological stations or from precipitation radar system. This evaluation procedure is performed similarily to the classification procedure described in step 1.

The main problem is to distinguish situations with enhanced natural radioactivity in the environment from artificial radioactivity due to accidental release. Thus, expert knowledge (level and duration of enhancement) for typical natural rain events is essential.

If no natural reason for significantly enhanced data exists, the staff member on duty has create an alarm in the IMIS system. In the past 16 years of operation typically 20 to 90 network internal alerts were generated per year but never the IMIS system has been changed to the emergency mode of operation.

Step 5: Emergency mode of operation

Authorised personnel (e.g. from governmental authorities) have to confirm the alarm. In this case, the IMIS information and monitoring systems are switched to the emergency mode of operation.

Experience from 16 years of operation

Since about 16 years the German gamma dose rate network is operated under more or less identical conditions with respect to the early warning mechanisms described above.

From figure 14 the yearly incidence of internal alerts over the past 16 years can be seen.

In the figure 15 the number of measuring stations involved in a network internal alert is shown. As can be seen, typically only 2 stations are generating an alert, but situations can also be found, where more than 150 stations show increased levels of radioactivity almost simultaneously. These situations occur typically only once a year during heavy thunder-storms.

In figure 16 the distribution of the duration of network internal alerts is shown. Most alerts have a duration of 2 to 6 hours. But the number on alerts with a duration of more than half a day can not be neglected. The staff members on duty frequently have to follow precipitation events which last more than even a day.

The last diagram shows the seasonal variation of internal alerts. During winter times network internal alerts occur only very seldom, but can be expected to take place from April to September.

Back to WP5 discussion.