Two surveys per spring are enough to obtain robust population trends of common and widespread birds in yearly monitoring programmes
Published: Volume 68.1, January 2021. Pages 33-51.
Extensive bird monitoring programmes are fundamental for estimating inter-annual population trends using data provided by thousands of observers through standardised fieldwork. Gordo (2018) has proposed that abundance data recorded by common bird monitoring schemes (e.g. SACRE programme) should be used cautiously due to its potential inaccuracy, because two surveys per spring are not enough to record the actual maximum number of individual birds at a sampling location. We carried out numerical simulations and analysed the interspecific pattern of statistical significance of the published population trends of the Spanish common birds census, the SACRE programme (1998-2011), in order to test how the number of repetitions of censuses per year affects the power of tests: (i.e. the probability of detecting significant trends that are in fact true), and the probability of obtaining low false discovery rates: i.e. identifying significant changes that are actually false, when estimating yearly population changes. we agree with Gordo (2018) that two surveys of the same sampling stations per year are unable to detect the maximum number of birds throughout a breeding season. Nevertheless, the goal of monitoring programmes is not to obtain the maximum number of birds at each sampling unit over a long time span but to measure reliable population trends. Our results demonstrate that the average number of birds recorded in two surveys per season provides a highly reliable indication of population trends for abundant and widely distributed bird species, the focal taxa in common birds monitoring schemes, especially of long-term average trends > ±2.5% change annually. The actual population trends for very rare species, such as those with data from fewer than 50 UTM squares and < 5 individual birds per census and UTM cell, are hard to detect unless they show yearly percentage population changes greater than ±5%.