L.L. Soldaat

Bias in phenology assessments based on first appearance data of butterflies

Data on the first appearance of species in the field season are widely used in phenological studies. However, there are probabilistic arguments for bias in estimates of phenological change if sampling methods or population abundances change. We examined the importance of bias in three measures of phenological change: (1) the date of the first X appearances, (2) the date of the first Y% of all first appearances and (3) the date of the first Z% of the individuals observed during the entire flight period. These measures were tested by resampling the data of the Dutch Butterfly Monitoring Scheme and by simulations using artificial data. We compared datasets differing in the number of sampling sites, population abundance and the start of the observation period. The date of the first X appearances proved to be sensitive to the number of sampling sites. Both the date of the first X appearances and the date of the first Y% of all first appearances were sensitive to population trend. No such biases were found for estimates of the first Z% of the flight period, but all three measures were sensitive to changes in the start of the observation period. The conclusions were similar for both the study on butterfly data and the simulation study. Bias in phenology assessments based on first appearance data may be considerable and should no longer be ignored in phenological research.

Smoothing and trend detection in waterbird monitoring data using structural time-series analysis and the Kalman filter

Many wildlife-monitoring programmes have long time series of species abundance that cannot be summarized adequately by linear trend lines. To describe long time series better, generalized additive models may be used to obtain a smooth trend line through abundance data. We describe another approach to estimate a smoothed trend line through time series consisting of one observation per time point, such as year or month. This method is based on structural time-series models in combination with the Kalman filter and is computerized in the TrendSpotter software. One of its strengths is the possibility to test changes in smoothed abundances between years, taking into account serial correlation. The trend method is applied in the Dutch Waterbird Monitoring Scheme (DWMS), a monitoring scheme for migrating and overwintering waterbirds. Taking the numbers of overwintering Greater Scaup (Aythia marila) in the Netherlands as an example, we demonstrate three applications of the method: (1) trend calculation and classification for each year in the time series, (2) assessing alerts for alarming population declines and (3) testing yearly abundance against a population threshold. We discuss the situations where TrendSpotter is to be preferred over other methods.

Occupancy dynamics of wild rabbits (Oryctolagus cuniculus) in the coastal dunes of the Netherlands with imperfect detection

Context Wild rabbits are considered a key species in the coastal dunes of the Netherlands, but populations have collapsed as a result of viral diseases. Aim We studied to what extent population collapse led to local extinction and whether recolonisation of empty patches in the dunes happened. Methods We investigated occupancy dynamics using data of 245 transects where rabbits were surveyed in 1984–2009. Dynamic site-occupancy models were used to analyse the data. These models adjust for imperfect detection to avoid bias in occupancy-trend estimation. Key results The decline of the rabbit population has resulted in many local extinctions, especially in woodland and in the northern part of the coastal dunes. Most transects along grassland and mixed vegetation have recently been reoccupied. The recovery of woodland occupancy is slow, probably not because of limited dispersal capacity of rabbits, but because the quality of woodland habitats is poor. Detection probability of rabbits varied considerably over the years and among habitat types, indicating the necessity of taking detection into account. Rabbits were slightly better detected when it was cloudy, windy and rainy and when lunar phase approached new moon. Conclusion Extinction and recolonisation of habitat patches varied considerably among habitat types. Implications The current slow recolonisation hampers the recovery of rabbit populations in woodland habitats in the Dutch coastal dunes. Furthermore, monitoring rabbit occupancy should take imperfect detection into account to avoid biased results.

Comment on “Working with population totals in the presence of missing data comparing imputation methods in terms of bias and precision” by Onkelinx et al. (2016)

In their recent paper, Onkelinx et al. (2016), hereafter called ONK16, present a novel application of multiple imputation (hereafter called MI) techniques to water bird censuses. This presentation is accompanied by a comparison of MI with two existing software packages for bird count analysis: UIndex (Underhill and Prŷs-Jones 1994) and BirdSTATs/TRIM (Pannekoek and van Strien 2005; van der Meij 2013). Although we fully agree that for some use cases, multiple imputation as a method is a useful alternative to analytical approaches, e.g., as used in TRIM, to infer the uncertainty associated with the analysis of trends and/or indices in bird count data, we do believe that the conclusions drawn about UIndex, BirdStats and TRIM are contingent upon a number of misconceptions about these programs. In this Comment, we summarize these, and make a brief assessment of their consequences, where appropriate.