Spatial Variation In Mortality Research Articles

Spatial memory predicts home range size and predation risk in pheasants.

Most animals confine their activities to a discrete home range, long assumed to reflect the fitness benefits of obtaining spatial knowledge about the landscape. However, few empirical studies have linked spatial memory to home range development or determined how selection operates on spatial memory via the latter's role in mediating space use. We assayed the cognitive ability of juvenile pheasants (Phasianus colchicus) reared under identical conditions before releasing them into the wild. Then, we used high-throughput tracking to record their movements as they developed their home ranges, and determined the location, timing and cause of mortality events. Individuals with greater spatial reference memory developed larger home ranges. Mortality risk from predators was highest at the periphery of an individual's home range in areas where they had less experience and opportunity to obtain spatial information. Predation risk was lower in individuals with greater spatial memory and larger core home ranges, suggesting selection may operate on spatial memory by increasing the ability to learn about predation risk across the landscape. Our results reveal that spatial memory, determined from abstract cognitive assays, shapes home range development and variation, and suggests predation risk selects for spatial memory via experience-dependent spatial variation in mortality.

Spatial variation of migrant‐native mortality differentials by duration of residence in Belgium: A story of partial convergence

AbstractPrevious studies have repeatedly shown that all‐cause mortality is subject to spatial variation within countries and that acculturation of migrant mortality to native patterns occurs gradually with increasing duration of residence. This suggests that spatial variation in mortality is likely to differ between migrants and natives and that the migrant mortality advantage may be subject to strong variation between local settings. Using longitudinal microdata from the Belgian census, the population register, and the tax registers for 2011–2015, this article contributes to our understanding of migrant‐native mortality differentials by considering subnational spatial variation in the mortality advantage, settlement patterns of migrant groups, and varying degrees of acculturation by duration of residence. Consistent with the health selection hypothesis and the migration as a rapid health transition hypothesis, spatial variation in all‐cause mortality is limited among migrants with durations of residence under 30 years, resulting in strong variation of the mortality advantage between districts. Partial convergence to the spatial pattern of natives emerges among first generation migrants with durations of residence exceeding 30 years, as well as the intermediate and second generation. In Belgium, the country‐level mortality advantage can be partially accounted for by the settlement pattern of the migrant population in districts characterised by high native mortality and large migrant‐native differentials in all‐cause mortality.

Bayesian Small Area Estimates of Diabetes Incidence by United States County, 2009

In the United States, diabetes is common and costly. Programs to prevent new cases of diabetes are often carried out at the level of the county, a unit of local government. Thus, efficient targeting of such programs requires county-level estimates of diabetes incidence-the fraction of the non-diabetic population who received their diagnosis of diabetes during the past 12 months. Previously, only estimates of prevalence-the overall fraction of population who have the disease-have been available at the county level. Counties with high prevalence might or might not be the same as counties with high incidence, due to spatial variation in mortality and relocation of persons with incident diabetes to another county. Existing methods cannot be used to estimate county-level diabetes incidence, because the fraction of the population who receive a diabetes diagnosis in any year is too small. Here, we extend previously developed methods of Bayesian small-area estimation of prevalence, using diffuse priors, to estimate diabetes incidence for all U.S. counties based on data from a survey designed to yield state-level estimates. We found high incidence in the southeastern United States, the Appalachian region, and in scattered counties throughout the western U.S. Our methods might be applicable in other circumstances in which all cases of a rare condition also must be cases of a more common condition (in this analysis, "newly diagnosed cases of diabetes" and "cases of diabetes"). If appropriate data are available, our methods can be used to estimate proportion of the population with the rare condition at greater geographic specificity than the data source was designed to provide.

The U‐shaped pattern of size‐dependent mortality and its correlated factors in a subtropical monsoon evergreen forest

Abstract Tree mortality is an important ecological process influencing multiple functions of forest ecosystems. Previous studies have shown two basic size–mortality patterns, including a competition‐driven declining and a disturbance‐driven increasing mortality rate with tree size. Subtropical forests, which have a high species diversity and subject to frequent monsoon disturbances, are widely distributed in eastern Asia. However, the tree size–mortality pattern in the mature subtropical forests remains unclear. Here we analysed patterns of size‐dependent mortality from tree species to forest community using a 5‐year inventory data from 117 species and 163,612 individuals in a 20‐ha forest dynamic plot in a mature subtropical monsoon evergreen forest in eastern China. To explain the spatial variability in mortality patterns, two major biotic drivers (competition and tree size) and multiple local‐scale environmental factors were further analysed. Our results showed that tree size was the best predictor of tree mortality at the scales of both species and community. A species‐level analysis identified four size–mortality patterns that are shaped by species‐specific attributes such as maximum size and life form. For 27 out of 92 species that comprised 59% of tree individuals, the relationship between size and mortality exhibited a U‐shaped pattern of a first decline followed by an increase. An overall community‐scale size‐dependent mortality also showed a U‐shaped pattern. Tree mortality was also influenced by the competition and environmental conditions, but the relative importance varied widely across tree sizes and species. The competition showed significant correlations with the mortality of small trees while the effect of environmental conditions on mortality was strongest for large trees. A principal component analysis showed that a combination of biotic and abiotic factors explained 42.3% of the spatial variation in mortality at large sizes. Synthesis. Our results reveal four identifiable size‐dependent mortality patterns that differ across diverse species, jointly leading to a U‐shaped size–mortality pattern at the community level. This finding calls for the need to establish the details of every potential size–mortality pattern with consideration of the different effects of biotic and abiotic factors on tree mortality of specific size.

Locally high, but regionally low: the impact of the 2014–2016 drought on the trees of semi-arid savannas, South Africa

Drought is thought to be a factor limiting the abundance of shrubs and trees in savannas, but this has rarely been tested empirically. The severe drought of 2014–2016 provided an opportunity to document the impact of drought on the mortality of woody plants in semi-arid savannas of South Africa. Annual data from long-term tree monitoring plots were combined with opportunistic sampling in areas of high mortality to determine the extent and severity of mortality across an area of approximately 10 000 km2. Although mortality was not extensive, certain landscapes suffered mortality similar to the highest values reported for semi-arid ecosystems. Rainfall deficit was identified as a key determinant of spatial variation in mortality at the regional scale, which varied from 3% to 49% over two years. Spatial variation at the landscape scale was also high, and affected primarily by species composition and height of the woody plants, with higher mortality for taller individuals. An effect of topographic position and unknown edaphic factors was also evident. Species that suffered high mortality included a common bush encroachment species, but also many non-encroachers. Additional study is needed to confirm that the observed mortality will cause a long-term reduction in woody plant density.

Reproductive strategy of a migratory fish stock: implications of spatial variations in natural mortality

We investigate how the reproductive strategy in a migratory marine fish may be influenced by spatial variations in mortality in early life stages. In particular, we examine how spawning time and location affect offspring survival and growth. A drift model for early life stages (eggs to age 1) of the Barents Sea cod (Gadus morhua) is combined with empirical estimates of spatial variation in mortality at two different life stages. We examine seasonal and interannual differences in survival and growth in offspring originating from two spawning grounds, with the central site requiring higher migration distance, and hence cost, than the northern site. When accounting for spatially explicit mortality fields, central and northern spawned offspring have about equal survival, as do early and late spawned offspring. Furthermore, central spawned offspring grow faster and are likely to reach a larger size compared with northern spawned offspring. Our results indicate that the fitness benefit of southward migration in the Barents Sea cod is not mainly due to higher early survival of offspring, but rather due to effects of offspring acquiring a larger size.

Spatial variation in mortality by in‐pot predation in the Tasmanian rock lobster fishery

AbstractThe capture of target species using pots or traps can be affected by interactions that occur inside fishing gear. In the Tasmanian southern rock lobster (Jasus edwardsii) fishery (SRLF), predation by octopus is known to vary spatially, assumed to be in response to different environmental and ecological conditions. However, these effects are not well understood and there is concern that changes in the system, as a result of climate change, may alter the spatial distribution and extent of octopus predation. Using fishery‐dependent commercial time series data (2000–2011), we first examined how in‐pot lobster mortality by octopus predation varied across fishing areas at inter‐ and intra‐annual scales. These spatial and temporal patterns were examined using dynamic regression models (DRM) using lagged variables of lobster abundance (lobster catch per unit effort, CPUE), fishing effort (potlifts) and environmental variables (sea surface temperature, SST). This demonstrated that in‐pot mortality in most areas was strongly affected by the current lobster CPUE (indicating an instantaneous direct impact), whereas fishing effort and SST differed spatially in terms of lag selection and correlation sign. Most eastern Tasmanian areas showed a stronger effect of fishing effort rather than SST, with the opposite pattern along the west coast. We provide the first classification of SRLF stock assessment areas based on predation risk level and identify the explanatory fishing and environmental factors.

Are strong fire–vegetation feedbacks needed to explain the spatial distribution of tropical tree cover?

AbstractAimThe spatial pattern of tropical fire‐induced tree mortality is partly determined by climate, but feedbacks of tree cover on fire are also important. We re‐examine some recent observations proposed as evidence for very strong tree‐cover feedbacks on fire, sufficient to allow savanna and forest to be alternative stable states over large areas of the tropics. Two pieces of previously reported observational evidence are examined: (1) the trimodal statistical distribution of tropical tree fraction, and (2) the fact that different tree fractional cover is found at different locations in the tropics with similar rainfall.LocationGlobal tropics.MethodsFor point (1) above we analyse the statistical distribution of tree fraction predicted by the logistic equation of tree growth and self‐competition, with spatially varying mortality rates. For (2), the relationship between mean annual rainfall and mean net primary productivity (NPP) in a climate model is examined.Results(1) A trimodal distribution of tree cover does not necessarily require tree‐cover feedback on fire. It can arise from a combination of two factors: nonlinearities in vegetation dynamics and climate‐driven spatial variation in mortality (the intermediate fire–productivity hypothesis). (2) Different locations in the tropics can have identical rainfall but significantly different NPP, even with no feedback of tree cover on fire.Main conclusionsOur results show that strong tree‐cover feedback on fire is not necessary to explain observations (1) and (2). However, it is still possible that strong fire–vegetation feedback is the primary explanation – our results do not rule this out. We simply demonstrate the possibility of an alternative hypothesis (of strong climate control). In reality, it is likely that both tree‐cover feedback and climate contribute. It is challenging to separate these two effects cleanly. More work is needed to quantify their separate effects. We show that plots of mortality versus productivity are useful tools for understanding spatial variations in tree cover.

Mapping Estimated Heat-Related Mortality In London Due To Population Age, Urban Heat Island, And Dwelling Characteristics

Introduction Mortality during hot weather is predicted to become more significant in the UK due to changes in the climate. We investigated the spatial variation in heat-related mortality risks across London due to population age, urban heat island (UHI) and building characteristics. Methods Dynamic building physics models were run for 126 different age and built form variants of buildings statistically representative of the London housing stock, outputting average maximum daily internal temperatures for a hot summer period in 2006. Internal and modelled UHI temperature estimates for the same period in London were mapped to individual dwellings in Geographic Information System (GIS) software, and deviations of building-specific indoor and outdoor temperatures from the London-average were calculated and summarised at ward level. Office of National Statistics age-specific absolute mortality data and the above-calculated UHI and indoor temperature anomalies were then used to estimate spatial variation in mortality according to an age-specific relative risk function for heat in London. Results There was appreciable variation in indoor temperatures under the modelled climate, with a mean maximum range of 2.5°C. Modelled results indicate that the largest mortality risks may lie in outer London, as the elderly population is larger there than in Central London. Top-floor flats in multiple occupancy buildings and bungalows were found to be the most at-risk dwelling types. Limitations in building stock information and occupancy scenario assumptions mean that some buildings may be worse than modelled. Conclusions These results indicate that UHI and indoor temperatures may cause an appreciable spatial variation in the risks from heat-related mortality, but that the driving factor remains population age. Further work is required to explore the risks for individuals living in extreme heat-poor housing conditions or for those unable to take action to cool the indoor environment.

Spatial variations in mortality in pelagic early life stages of a marine fish (Gadus morhua)

Mortality of pelagic eggs and larvae of marine fish is often assumed to be constant both in space and time due to lacking information. This may, however, be a gross oversimplification, as early life stages are likely to experience large variations in mortality both in time and space. In this paper we develop a method for estimating the spatial variability in mortality of eggs and larvae. The method relies on survey data and physical–biological particle-drift models to predict the drift of ichthyoplankton. Furthermore, the method was used to estimate the spatially resolved mortality field in the egg and larval stages of Barents Sea cod (Gadus morhua). We analyzed data from the Barents Sea for the period between 1959 and 1993 when there are two surveys available: a spring and a summer survey. An individual-based physical–biological particle-drift model, tailored to the egg and larval stages of Barents Sea cod, was used to predict the drift trajectories from the observed stage-specific distributions in spring to the time of observation in the summer, a drift time of approximately 45days. We interpreted the spatial patterns in the differences between the predicted and observed abundance distributions in summer as reflecting the spatial patterns in mortality over the drift period. Using the estimated mortality fields, we show that the spatial variations in mortality might have a significant impact on survival to later life stages and we suggest that there may be trade-offs between increased early survival in off shore regions and reduced probability of ending up in the favorable nursing grounds in the Barents Sea. In addition, we show that accounting for the estimated mortality field, improves the correlation between a simulated recruitment index and observation-based indices of juvenile abundance.

0204 Mesoamerican nephropathy in Costa Rica: Geographical distribution and time trends of chronic kidney disease mortality between 1970 and 2012

ObjectivesTo characterise geographical distribution and time trends of chronic kidney disease (CKD) mortality in the context of the epidemic of Mesoamerican nephropathy (MeN), likely related to occupational heat stress and...

Spatial variations in motorcycle registrations and the mortality of motorcycle users due to traffic injuries in Argentina.

Although pedestrians, cyclists and motorcyclists account for nearly half of those killed in traffic accidents in the world, little is known about the geographical distribution patterns of these vulnerable roadway users. Using spatial scan statistics techniques, the spatiotemporal variations in the mortality of motorcycle users in Argentina were analyzed for the period 2001-2010, as well as the spatial variations in mortality and motorcycle registration from 2007-2010. Two space-time clusters with a high risk of death for motorcycle riders were identified during the second half of the study period. Overall, there was a spatial relationship between motorcycle registrations and the mortality of the users of these vehicles in the northern-central region of Argentina. The results of this study indicate the need to reinforce primary prevention policies focused on motorcycle users in this region of the country, especially in areas with high population density.

Spatial variations in motorcycle registrations and the mortality of motorcycle users due to traffic injuries in Argentina

Although pedestrians, cyclists and motorcyclists account for nearly half of those killed in traffic accidents in the world, little is known about the geographical distribution patterns of these vulnerable roadway users. Using spatial scan statistics techniques, the spatiotemporal variations in the mortality of motorcycle users in Argentina were analyzed for the period 2001-2010, as well as the spatial variations in mortality and motorcycle registration from 2007-2010. Two space-time clusters with a high risk of death for motorcycle riders were identified during the second half of the study period. Overall, there was a spatial relationship between motorcycle registrations and the mortality of the users of these vehicles in the northern-central region of Argentina. The results of this study indicate the need to reinforce primary prevention policies focused on motorcycle users in this region of the country, especially in areas with high population density.

Window Area and Development Drive Spatial Variation in Bird-Window Collisions in an Urban Landscape

Collisions with windows are an important human-related threat to birds in urban landscapes. However, the proximate drivers of collisions are not well understood, and no study has examined spatial variation in mortality in an urban setting. We hypothesized that the number of fatalities at buildings varies with window area and habitat features that influence avian community structure. In 2010 we documented bird-window collisions (BWCs) and characterized avian community structure at 20 buildings in an urban landscape in northwestern Illinois, USA. For each building and season, we conducted 21 daily surveys for carcasses and nine point count surveys to estimate relative abundance, richness, and diversity. Our sampling design was informed by experimentally estimated carcass persistence times and detection probabilities. We used linear and generalized linear mixed models to evaluate how habitat features influenced community structure and how mortality was affected by window area and factors that correlated with community structure. The most-supported model was consistent for all community indices and included effects of season, development, and distance to vegetated lots. BWCs were related positively to window area and negatively to development. We documented mortalities for 16/72 (22%) species (34 total carcasses) recorded at buildings, and BWCs were greater for juveniles than adults. Based on the most-supported model of BWCs, the median number of annual predicted fatalities at study buildings was 3 (range = 0–52). These results suggest that patchily distributed environmental resources and levels of window area in buildings create spatial variation in BWCs within and among urban areas. Current mortality estimates place little emphasis on spatial variation, which precludes a fundamental understanding of the issue. To focus conservation efforts, we illustrate how knowledge of the structural and environmental factors that influence bird-window collisions can be used to predict fatalities in the broader landscape.

Recent climate changes interact with stand structure and management to determine changes in tree carbon stocks in Spanish forests

AbstractMost temperate forests are accumulating carbon (C) and may continue to do so in the near future. However, the situation may be different in water‐limited ecosystems, where the potentially positive effects of C and N fertilization and rising temperatures interact with water availability. In this study, we use the extensive network of plots of two consecutive Spanish national forest inventories to identify the factors that determine the spatial variation of the C stock change, growth, and mortality rate of forests in Peninsular Spain (below‐ and aboveground). We fitted general linear models to assess the response of C stock change and its components to the spatial variability of climate (in terms of water availability), forest structure (tree density and C stock), previous forest management, and the recent warming trend. Our results show that undisturbed forests in Peninsular Spain are accumulating C at a rate of ~1.4 Mg C ha−1 yr−1, and that forest structural variables are the main determinants of forest growth and C stock change. Water availability was positively related to growth and C accumulation. On the other hand, recent warming has reduced growth rate and C accumulation, especially in wet areas. Spatial variation in mortality (in terms of C loss) was mostly driven by differences in growth rate across plots, and was consistent with ‘natural’, self‐thinning dynamics related to the recent abandonment of forest management over large areas of Spain, with the consequent increase in tree density and competition. Interestingly, the negative effect of warming on forest C accumulation disappears if only managed stands are considered, emphasizing the potential of forest management to mitigate the effects of climate change. However, the effect of forest management was weak and, in some cases, not significant, implying the need of further research on its impact.

How much variation in tree mortality is predicted by soil and topography in Central Amazonia?

Tree mortality in Amazonia has been related to regional variation in soil, topography and climatic disturbances, but the magnitude of the effect of these factors on tree mortality at local and mesoscales remains poorly determined. We investigated tree mortality in 72 1-ha permanent plots spanning 64 km 2 of tropical moist forest in Reserva Ducke, Manaus, Brazil. Plots were censused three times (2000–2003, 2003–2005, and 2005–2008), resulting in two census intervals. The relationships of soil and topography to tree mortality were dependent on tree size. Small- and medium-sized trees (1 ⩽ dbh < 30 cm) had similar relationships of mortality with soil and topography, while large trees (dbh ⩾ 30 cm) showed different (or no) relationships. The effects of soil and topography on tree mortality also varied temporally. In the second census interval after storms, soil and topography explained about one-fourth of the spatial variation in mortality of small- and medium-sized trees (<30 cm dbh), whereas no effects were detected in the first census interval. In particular, soil fertility was the most important predictor of tree mortality in the study area. Topography alone (altitude and slope) was associated with only 12% of the spatial variation in tree mortality and the magnitude of the effect of soil and topography on tree mortality also increased after storms. In general, plots on more fertile soils, on steep slopes and sandy soils in valleys showed greater tree mortality than those on plateau with well-drained clayey soils. Therefore, disturbance history and tree size should be included when scaling up tree mortality from local to regional scales. As much variation remains unexplained, other landscape features, such as watershed morphology and wind exposure, may be necessary to make more precise predictions on patterns of tree mortality in Central Amazonia.

Spatial analysis of heat-related mortality among the elderly between 1993 and 2004 in Sydney, Australia

This study analyzed the geographical patterns of heat-related mortality among the population aged 65 and over within the metropolitan area of Sydney, Australia between 1993 and 2004, and evaluated the role of some physical and socio-demographic risk factors associated with it. The effect of temperature on all-cause mortality during unusually hot days was investigated using spatial analytic techniques, such as cluster analysis and spatial regression analysis. Generalized Linear Models (GLMs) were used to investigate the role of daily average temperature, ozone (O(3)) and particulate matter of diameter less than 10 microm (PM(10)) at the regions that showed a significant increase in mortality on unusually hot days. Spatial variation in mortality on unusually hot days was observed among the population 65 and over. Elderly people living within 5-20 km south-west and west of the Sydney Central Business District (CBD) were found to be more vulnerable. However, analysis using GLMs showed temperature to be a significant modifier of daily mortality in the region to the south-west of the CBD only. O(3) and PM(10) were found to be non-significant factors in the regions where air pollutants were studied. Socio-economic status and the proportion of vegetation or developed land in each Statistical Local Area (SLA) were also not a significant factor explaining the increased mortality. A combination of social and environmental factors may be at play. Our results suggest an effect of temperature on mortality of the elderly population in Sydney Statistical Division at the SLA level. More spatially-based research would be beneficial once climate datasets with improved spatial coverage become available.

Exploring Spatial Structure behind the Road Mortality of Regions in Europe

The Bayesian intrinsic conditional autoregressive convolution model was used to study the spatial variations in road mortality in the regions of Belgium, France and Germany. In all three countries, the spatial structure behind relative risk is significant and spatial heterogeneity predominates over unstructured heterogeneity. The maps of spatially-structured component of random effect enable the spatial structures of risk to be identified and highlight the zones and areas where the mortality risk exhibits spatial dependency. Hence, a west–east gradient in risk level is found in Germany and a north–south gradient in Belgium. In areas with high road network density (and relatively high population density), there is generally less heterogeneity in road mortality across neighbouring regions, while in areas with high regional disparities, there are significant spatial variations in mortality risk. Furthermore, a model was produced for 272 regions in 13 European continental countries, making it possible to investigate whether national borders have any specific effect on the distribution of road mortality risk compared to what happens within countries’ administrative borders. Cross-border regions were found not to be particularly likely to have similar road mortality risk levels where they shared a common national border. National borders have no specific effect on the distribution of road mortality across Europe.

Neither the devil nor the deep blue sea: larval mortality factors fail to explain the abundance and distribution ofTischeria ekebladella

Abstract1. The distribution, abundance and population dynamics of herbivorous insects may be affected by trophic interactions, by abiotic influences, or by intra‐specific processes. Relatively little is known about how trophic influences vary across space. Here, we investigate spatial variation in mortality in the oak‐feeding leaf minerTischeria ekebladellaas attributable to individual causal agents.2. Leaf miners were experimentally introduced on 67 trees on an island 5 km2in area in south‐western Finland. On each tree, some larvae were protected by a muslin bag, others by a glue barrier around the branch and some left exposed.3. In the bagged transplants, 78.4% of larvae survived, compared with only 9.6% in the other two treatments. Most of the mortality was because of airborne agents: mortality on branches sheltered by a glue barrier was as high as on fully exposed branch tips.4. We consider mortality caused by parasitoid wasps to be the main source of larval death and the primary factor driving general patterns of survival. The effects of bird predation and premature leaf abscission were negligible.5. We detected spatial aggregation in larval survival and parasitism rates at the level of individual trees, but not across the landscape.6. Spatial variation in overall leaf miner survival, parasitism and leaf abscission does not suffice to explain patterns of incidence and abundance of wildT. ekebladellaon experimental trees. Rather, we identify metapopulation dynamics as a likely determinant of the spatial distribution ofT. ekebladellain the landscape.

SPATIO-TEMPORAL CHANGES IN MORTALITY, GROWTH AND CONDITION OF THE PACIFIC OYSTER, CRASSOSTREA GIGAS, IN NORMANDY (FRANCE)

Summer mortalities have been observed in French shellfish areas (including Normandy) since 1994, but origin of this syndrome remains unclear and is suspected to be caused by a combination of several interacting extrinsic (biotic and abiotic) and intrinsic (genetic, physiological, immunological) factors. The French research program, MOREST aimed to identify the origin of oyster summer mortality along the French coast, focusing on the interactions between oysters, their pathogens, and the environment. The present study analyzed spatio-temporal variation in growth, condition and mortality in spat, and half-grown and market-sized oysters reared from February 2000 to October 2003 at six stations within two different areas in the Bay of Veys, Normandy: Grandcamp (GR), and Gefosse (GE) that is more estuarine. These biological parameters were compared between years, age groups, and areas. Results showed that shell growth was significantly lower in the station highest on the shore and similar in the four other sites, whereas tissue growth and condition index were higher in the Gefosse area. Results also showed large interannual, interage, seasonal and spatial variation in oyster mortality. In 2001, mortalities were markedly higher than in other years and all batches and sites were affected by high mortality rates. Moreover oysters suffered much higher mortalities in their second and third years than as spat, and the difference between age classes was accentuated when mortality was high. Increases in mortality occurred when the gonad was most extensively developed and the peak coincided with the spawning and postspawning periods when gonad volume began to decrease. Spatial variability showed that mortality was higher in Gefosse than Grandcamp. Chronology and spatial variations in mortality highlighted the importance of timing and confirmed that mass mortalities may be closely linked to reproduction. The risk seems to be associated with high reproductive effort, partial spawning, and/or slow gonad resorption. Spatial variation also suggests that the study sites experience varying degrees of stress caused by biological changes and probable differences in water quality reflecting the influence of freshwater input in Gefosse where mortality is higher.