The color of greater flamingo feathers fades when no cosmetics are applied.

Greater flamingos use cosmetic coloration by spreading uropygial secretions pigmented with carotenoids over their feathers, which makes the plumage redder. Because flamingos inhabit open environments that receive direct solar radiation during daytime, and carotenoids bleach when exposed to solar radiation, we expected that the plumage color would fade if there is no maintenance for cosmetic purposes. Here, we show that the concentrations of pigments inside feathers and on the surface of feathers were correlated, as well as that there was a correlation between the concentrations of pigments in the uropygial secretions and on the surface of feathers. There was fading in color (becoming less red) in feathers that received direct solar radiation when there was no plumage maintenance, but not so in others maintained in darkness. When we controlled for the initial color of feathers, the feathers of those individuals with higher concentration of pigments on the feather surfaces were those that lost less coloration after experimental exposure of feathers to sunny conditions. These results indicate that exposure to sunlight is correlated with the fading of feather color, which suggests that individuals need to regularly apply makeup to be more colorful. These results also reinforce the view that these birds use cosmetic coloration as a signal amplifier of plumage color. This may be important in species using highly variable habitats, such as wetlands, since the conditions experienced when molting may differ from those when the signal should be functional, usually months after molting.

Combining correlative and mechanistic habitat suitability models to improve ecological compensation.

Only a few studies have shown positive impacts of ecological compensation on species dynamics affected by human activities. We argue that this is due to inappropriate methods used to forecast required compensation in environmental impact assessments. These assessments are mostly descriptive and only valid at limited spatial and temporal scales. However, habitat suitability models developed to predict the impacts of environmental changes on potential species' distributions should provide rigorous science-based tools for compensation planning. Here we describe the two main classes of predictive models: correlative models and individual-based mechanistic models. We show how these models can be used alone or synoptically to improve compensation planning. While correlative models are easier to implement, they tend to ignore underlying ecological processes and lack accuracy. On the contrary, individual-based mechanistic models can integrate biological interactions, dispersal ability and adaptation. Moreover, among mechanistic models, those considering animal energy balance are particularly efficient at predicting the impact of foraging habitat loss. However, mechanistic models require more field data compared to correlative models. Hence we present two approaches which combine both methods for compensation planning, especially in relation to the spatial scale considered. We show how the availability of biological databases and software enabling fast and accurate population projections could be advantageously used to assess ecological compensation requirement efficiently in environmental impact assessments.

Impacts of extreme climatic events on the energetics of long-lived vertebrates: the case of the greater flamingo facing cold spells in the Camargue.

Most studies analyzing the effects of global warming on wild populations focus on gradual temperature changes, yet it is also important to understand the impact of extreme climatic events. Here we studied the effect of two cold spells (January 1985 and February 2012) on the energetics of greater flamingos (Phoenicopterus roseus) in the Camargue (southern France). To understand the cause of observed flamingo mass mortalities, we first assessed the energy stores of flamingos found dead in February 2012, and compared them with those found in other bird species exposed to cold spells and/or fasting. Second, we evaluated the monthly energy requirements of flamingos across 1980-2012 using the mechanistic model Niche Mapper. Our results show that the body lipids of flamingos found dead in 2012 corresponded to 2.6±0.3% of total body mass, which is close to results found in woodcocks (Scolopax rusticola) that died from starvation during a cold spell (1.7±0.1%), and much lower than in woodcocks which were fed throughout this same cold spell (13.0±2%). Further, Niche Mapper predicted that flamingo energy requirements were highest (+6-7%) during the 1985 and 2012 cold spells compared with 'normal' winters. This increase was primarily driven by cold air temperatures. Overall, our findings strongly suggest that flamingos starved to death during both cold spells. This study demonstrates the relevance of using mechanistic energetics modelling and body condition analyses to understand and predict the impact of extreme climatic events on animal energy balance and winter survival probabilities.

Non-linear feeding functional responses in the Greater Flamingo (Phoenicopterus roseus) predict immediate negative impact of wetland degradation on this flagship species.

Accurate knowledge of the functional response of predators to prey density is essential for understanding food web dynamics, to parameterize mechanistic models of animal responses to environmental change, and for designing appropriate conservation measures. Greater flamingos (Phoenicopterus roseus), a flagship species of Mediterranean wetlands, primarily feed on Artemias (Artemia spp.) in commercial salt pans, an industry which may collapse for economic reasons. Flamingos also feed on alternative prey such as Chironomid larvae (e.g., Chironomid spp.) and rice seeds (Oryza sativa). However, the profitability of these food items for flamingos remains unknown. We determined the functional responses of flamingos feeding on Artemias, Chironomids, or rice. Experiments were conducted on 11 captive flamingos. For each food item, we offered different ranges of food densities, up to 13 times natural abundance. Video footage allowed estimating intake rates. Contrary to theoretical predictions for filter feeders, intake rates did not increase linearly with increasing food density (type I). Intake rates rather increased asymptotically with increasing food density (type II) or followed a sigmoid shape (type III). Hence, flamingos were not able to ingest food in direct proportion to their abundance, possibly because of unique bill structure resulting in limited filtering capabilities. Overall, flamingos foraged more efficiently on Artemias. When feeding on Chironomids, birds had lower instantaneous rates of food discovery and required more time to extract food from the sediment and ingest it, than when filtering Artemias from the water column. However, feeding on rice was energetically more profitable for flamingos than feeding on Artemias or Chironomids, explaining their attraction for rice fields. Crucially, we found that food densities required for flamingos to reach asymptotic intake rates are rarely met under natural conditions. This allows us to predict an immediate negative effect of any decrease in prey density upon flamingo foraging performance.