Community changes in odonate monitoring: why are long‐term studies so relevant?

Most ecological studies involving insects are based on medium‐ and short‐term observations; however, the extent to which such data captures reality remains unclear. We investigated the long‐term dynamics of two Odonata communities (disturbed and undisturbed sites) over 18 years and analysed the differences in the short‐ and long‐term results. We also focused on the sampling methodology to enhance the efficacy and objectivity of long‐term monitoring involving Odonata. During one year, we captured only 53% of the overall species richness; during three consecutive sampling years, it was 65%. To capture 95%, we needed 16 years. Changes in quantitative similarity (Renkonen index, P) were more pronounced within sites over time than between sites. Species constancy significantly increased with the maximum abundance class but decreased with increasing fluctuation ratio and specialisation (Dragonfly Biotic Index). Based on exuviae, we detected half of the species compared to adults, but the species accumulation curves peaked after a few sampling years. Long‐ and short‐term monitoring yield different results, both qualitatively (species richness, specialisation) and quantitatively (abundance, dominance). Ideal sampling should be sequential, lasting at least 10 years (capturing >80% of species). Intermittent sampling (one‐year interspersed with pauses), allowing the inclusion of multiple sites in monitoring programme, may also provide satisfactory results when performed over a longer period. Over the long term, sampling adults semi‐quantitatively and exuviae qualitatively provided sufficient information, while being feasible in terms of both personnel and costs, thereby overcoming the main pitfalls of long‐term monitoring programmes. Long‐ and short‐term monitoring yield different results, both qualitatively (species richness, specialisation) and quantitatively (abundance, patterns of dominance). During one year, we captured only 53% of the overall species richness. Ideal sampling should be sequential, lasting at least 10 years (capturing >80% of species). Intermittent sampling (one‐year interspersed with pauses) may provide satisfactory results when performed over a longer period. Sampling adults semi‐quantitatively and exuviae qualitatively provided sufficient information, while being feasible in terms of both personnel and costs, thereby overcoming the main pitfalls of long‐term monitoring programmes.