3. RESULTS
Regarding the 12 models of random effect structure, the model with no random effects structure and COM-Poisson error distribution was selected as the most parsimonious one (Table S1 in Supplementary material).
Regarding the 16 alternatives fixed effects models (all including the permanent variable REGION), the model with the lowest AICc included also AGRI5 and crop variables, SHANG and BARE (Table 1).
Compared with the North Ostrobothnia region, which showed stable or slightly positive population growth, all other regions were associated with strongly negative rates of population change (Table 2, Figure 3). Crop type diversity (SHANG) was, as expected, associated with positive rates of change (Table 2, Figure 4). Considering the proportion of farmland within the landscape surrounding the singing group (AGRI5) and the proportion of bare ground within the singing group’s range (BARE), the coefficient point estimates show a positive association with the rate of change but do not differ clearly from zero (Table 2, Figure 4).
In the whole of Finland, the average logarithmic rate of population growth was –0.21 (SE = 0.04; excluding Central Finland and North Carelia which had so low sample sizes).
4. DISCUSSION
Our results showed that the multiplicative population rate of change of Ortolan Buntings within the study period from 2000 to 2018 was positively associated with greater crop type diversity and with higher proportion of bare ground within a singing group’s range. Also, the proportion of farmland surrounding the singing group had a positive association with the rate of population change. Notably, the mean rate of population growth was positive only in the North Ostrobothnia region, while all other regions showed negative growth rates.
4.1 Importance of crop type diversity and bare ground
Our study revealed a positive association between Ortolan Bunting population growth rate and the two crop variables, crop type diversity and proportion of bare ground. Albeit the association was rather weak for bare ground. High crop type diversity at the territory scale may emerge as a critical factor, as single crop types rarely offer suitable vegetation structure for both nesting and foraging throughout the breeding season. Ground-feeding birds, like the Ortolan Bunting, may face challenges in finding food (insects, arthropods etc.) within impenetrable, tall and dense vegetation (Schifferli, 2001) usually associated with intensive farming. On the other hand, various taller or denser vegetation is needed for providing safe nest sites. Therefore, a mosaic of diverse crop plants of different sowing times, sward heights and densities as well as patches of bare ground may better support the Ortolan Bunting which may otherwise suffer from higher costs of reproduction, reduced breeding success and chick survival.
The importance of these crop variables has been found in earlier habitat preference studies as well. In central Sweden, Ortolan Buntings were found to prefer habitats which had heterogeneous ground vegetation characterized by patches with bare ground, or at least sparse ground vegetation, intermixed with patches with taller vegetation (Berg, 2008). In Finland, in the Tavastia region, Vepsäläinen et al. (2005) showed that the area of field growing spring cereal, root crops, etc. (i.e., field not covered in springtime) had a positive effect on the occurrence of the Ortolan Bunting, although they witnessed a population crash happening in spite of an increase in the area of non-covered field. Santana et al. (2017) however, found that in open Mediterranean farmland, when considering bird species richness, it was more important what crops were grown than their diversity or spatial configuration. Also Morelli (2012) showed that the occurrence of Central Italian Ortolan Buntings was not clearly related to structural heterogeneity, rather the species seemed to be associated with sunflower fields, oats and alfalfa. Yet, the authors concluded that these crop types may be preferred because of certain structural characteristics that may increase space with bare soil, which would be consistent with our findings.
Climate change might affect, now and in the future, the choice between grains that are sown in the spring or in the autumn. If farmers choose to grow more autumn sown grains, the proportion of bare ground in the spring, at the time of feeding the nestlings, will reduce in the future. This change will most likely affect Ortolan Buntings negatively, starting from the southwestern populations which occupy areas that are specialized in growing grains. Similarly, the increasing popularity of direct sowing techniques might be reducing the amount of ploughed field areas i.e., bare ground, affecting Ortolan Buntings negatively.