Monitoring of insects in veteran oaks

One hundred oak trees were selected from a set of 600 veteran oaks registered in the national oak survey. These oaks were divided into two sets of 50 oaks, where insects can be sampled every second year in each oak. For practical reasons, these two sets represent two different geographical regions; region East (Oslo, Akershus, Buskerud, Østfold, Telemark & Vestfold) and region West (Agder, Rogaland og Vestland). The oaks for each set were selected at random, with the exception of a doubled likelihood of selecting oaks with a visible cavity in addition to diameter at breast height larger than 200 cm, and a few criteria relating to the number of trees in each plot. The distribution of oaks between counties and within/outside of forest was controlled after the selection was made and was considered adequately balanced. Insects were sampled from the 50 selected oaks in region East in June and July in 2023,  50 selected oaks in the Western region in June - September 2024, and again from the same 50 selected oaks in region East between April and August in 2025. Two window traps were mounted at each tree, one in the canopy and one at the trunk, next to a cavity if the tree had any. Beetles were identified.

Several original tables are here combined into a single event table, to fit the GBIF event core data format. These tables together represent the sampling design of the monitoring program and are noted as a series of hierarchical event levels in the "samplingProtocol" column.The data should therefore be unpacked to make proper sense of the data structure.

Brief explanation of the hierarchical structure of the dataset:

1) The occurrence table can be joined to the event table through the parentEvent, which joins to an
2) identification event. This level exists because any sample may have gone through several identification events, possibly with differing methods. The identification events joins through its parentEvent with
3) a sampling_trap event, which designates a single trap in a single sampling event at a location. Sampling_trap events are joined through their parentEvent to a
4) locality sampling event, which is a single sampling period in a locality. Sampling events can have 1 or more traps (sampling trap events). Finally, the locality sampling events can be joined through their parentEvent to a
5) year locality event, which designates the sampling of insects in a single locality in a year.

Relevant metadata or collected explanatory data is attached to each level, with the dynamicProperties column collecting the datatypes that the Darwin Event Core doesn't presently cater to. For example a range of environmental data at the sampling sites.

An R-script for unpacking the data into a more usable format is available at https://github.com/NINAnor/national_insect_monitoring. Users will have to make minor adjustments to the script to fit this particular data set.