As field determinations take much effort, it would be useful to be able to predict easily the coefficients describing the functional response of free-living predators, the function relating food intake rate to the abundance of food organisms in the environment. As a means easily to parameterise an individual-based model of shorebird Charadriiformes populations, we attempted this for shorebirds eating macro-invertebrates. Intake rate is measured as the ash-free dry mass (AFDM) per second of active foraging ; i.e. excluding time spent on digestive pauses and other activities, such as preening. The present and previous studies show that the general shape of the functional response in shorebirds eating approximately the same size of prey across the full range of prey density is a decelerating rise to a plateau, thus approximating the Holling type II (' disc equation ') formulation. But field studies confirmed that the asymptote was not set by handling time, as assumed by the disc equation, because only about half the foraging time was spent in successfully or unsuccessfully attacking and handling prey, the rest being devoted to searching.A review of 30 functional responses showed that intake rate in free-living shorebirds varied independently of prey density over a wide range, with the asymptote being reached at very low prey densities (<150/m x2 ). Accordingly, most of the many studies of shorebird intake rate have probably been conducted at or near the asymptote of the functional response, suggesting that equations that predict intake rate should also predict the asymptote.A multivariate analysis of 468 ' spot ' estimates of intake rates from 26 shorebirds identified ten variables, representing prey and shorebird characteristics, that accounted for 81% of the variance in logarithm-transformed intake rate. But four-variables accounted for almost as much (77.3%), these being bird size, prey size, whether the bird was an oystercatcher Haematopus ostralegus eating mussels Mytilus edulis, or breeding. The four variable equation under-predicted, on average, the observed 30 estimates of the asymptote by 11.6 %, but this discrepancy was reduced to 0.2% when two suspect estimates from one early study in the 1960s were removed. The equation therefore predicted the observed asymptote very successfully in 93% of cases.We conclude that the asymptote can be reliably predicted from just four easily measured variables. Indeed, if the birds are not breeding and are not oystercatchers eating mussels, reliable predictions can be obtained using just two variables, bird and prey sizes. A multivariate analysis of 23 estimates of the half-asymptote constant suggested they were smaller when prey were small but greater when the birds were large, especially in oystercatchers. The resulting equation could be used to predict the half-asymptote constant, but its predictive power has yet to be tested.As well as predicting the asymptote of the functional response, the equations will enable research workers engaged in many areas of shore...

Portugal is building a land cover monitoring system to deliver land cover products annually for its mainland territory. This paper presents the methodology developed to produce a prototype relative to 2018 as the first land cover map of the future annual map series (COSsim). A total of thirteen land cover classes are represented, including the most important tree species in Portugal. The mapping approach developed includes two levels of spatial stratification based on landscape dynamics. Strata are analysed independently at the higher level, while nested sublevels can share data and procedures. Multiple stages of analysis are implemented in which subsequent stages improve the outputs of precedent stages. The goal is to adjust mapping to the local landscape and tackle specific problems or divide complex mapping tasks in several parts. Supervised classification of Sentinel-2 time series and post-classification analysis with expert knowledge were performed throughout four stages. The overall accuracy of the map is estimated at 81.3% (±2.1) at the 95% confidence level. Higher thematic accuracy was achieved in southern Portugal, and expert knowledge significantly improved the quality of the map.

A review of the scientific literature and an analysis of unpublished material identified ~ 50 possible alien molluscs from Greece, four of which were newly reported here. Records of ~ 100 additional taxa, which would strongly inflate the alien species numbers, were excluded. Among the ~ 50 candidate species, 43 were confirmed as alien and 12 as cryptogenic. Twenty-nine alien species were considered established, and four deemed invasive. Our results are consistent with the position of Greece in the east-west Mediterranean gradient, as well as the Aegean marine environment. The contribution of well-informed citizen scientists appears to be crucial to our overall knowledge of alien molluscan biotas because only 12 alien taxa were detected during formal research projects, while 31 were first found by amateurs. No molluscan introductions were confirmed from Greece pre-1960s, and subsequent periods had variable numbers of introductions, ranging from 0 (1976-1980) to 9.5 (2006-2010). The areas with the highest alien species concentrations were the Saronikos Gulf, near the Piraeus port area, and the Dodekanisa, near the Levantine coastline. Despite a general decline in taxonomic expertise and in local knowledge, we show that these are still needed when compiling and analyzing alien species inventories that subsequently influence policy and management decisions.

No abstract

Classification accuracy of remote sensing images with supervised learning depends on the quality and characteristics of training samples. Size is a key aspect of a sample and its impact on classification depends on several factors, including the classifier employed, dimension on the feature space and land cover characteristics. Random Forest classifier is considered to be of low sensitivity to variations in sample size. However, further investigation is required when feature spaces are large and training is performed with spectral subclasses of the land cover classes to be mapped. This paper proposes to assess the impact of sample size in the classification accuracy of Random Forest using multitemporal Sentinel-2 data and a detailed set of training subclasses to produce a map with general land cover classes. The results revealed similar classification accuracies after major reductions in sample size.