The observational systems utilised during the TRACT experimental campa
ign (September 1992) can roughly be divided into: existing climate sta
tions and towers: background chemical monitoring stations; aerological
measurements; instrumented aircraft; and special micrometeorological
stations operated only during the experimental period. Database 4, whi
ch contains the latter surface measurements, includes station descript
ions and data from 21 sites operated by nine different groups during t
he campaign period. These data will be described briefly. There are a
number of advantages in using a relational database for analysis of th
is data. Without creating an unnecessarily large database which is awk
ward to view and slow to process, specific measurements can be selecte
d easily since data from each site have been standardized using a comm
on time index. This allows comparison of the same or different variabl
es from one or more sites for investigating data quality or the effect
s of site location. Data can readily be combined into hourly or daily
averages or time series for comparison with other data sets (e.g. mode
l or aircraft data). Spatial analysis is also straightforward, althoug
h the clustering of the sites in the centre of the investigation area
means that care must be taken. Some preliminary comparisons are made h
ere in order to demonstrate the availability and reliability of the da
ta.