Developing a non-invasive indicator of pinniped health : Neonate behavior and growth in California sea lions ( Zalophus californianus )

Assessing the health of wildlife populations is critical to achieving conservation goals; however, proper assessments can be complicated when study sites are difficult to reach or when focal species are sensitive to human disturbance. Condition and growth of offspring may indicate population health, but obtaining such data generally relies on invasive techniques. Here, we examine the extent to which noninvasive observations of neonate behaviors could serve as a proxy for traditional approaches to estimating neonate body condition and growth of a wild pinniped, the California sea lion (Zalophus californianus), in the Gulf of California, Mexico. Generalized linear models and multiple linear regressions were used to examine the effect of sex, breeding island, year, and neonate behaviors (e.g., nursing, active) on body condition and growth rates. We found a strong correlation between individual growth rates and behaviors of male neonates. Males engaged in proportionally more active behaviors had lower growth rates ( = –0.0005), whereas males engaged in proportionally more nursing events had higher growth rates ( = 0.002). There was no relationship between neonate behavior and condition for either sex, nor between behavior and growth rates of female pups. These results provide insight into the possibility of using behavior as an indicator of individual status of males that could facilitate assessments of population status and illustrate the importance of using alternative approaches to measure animal condition for species conservation.


INTRODUCTION
Neonate condition may be a useful indicator of population health in mammals, yet most studies focus on general demographic indicators such as population size and vital rates (White andLubow 2002, McMahon et al. 2005).While numerical counts of offspring born within a given year explain proximate population health, the growth and body condition of offspring may provide a tool to predict future trends.Here, we use the term population health to define the proximate and long-term livelihood of an isolated or interconnected population, and body condition as the current state affecting short-and long-term fitness of an individual within a population.More in-depth details of neonate condition (e.g., growth rates) may enhance our understanding and predictive capacities of population trends because the condition of individuals is directly related to the population to which they belong (Unsworth et al. 1999, Pojar andBowden 2004).
Body size and condition influence the survival of all age classes (Clutton-Brock et al. 1997, Boltnev et al. 1998), often with the largest impact on neonates and juveniles because of their inexperience in searching for prey and their undeveloped immune systems (Singer et al. 1997).The condition of neonates at birth may be influenced by maternal factors (e.g., age, foraging experience, body size, and condition), along with the quality of the birth colony (Hastings andTesta 1998, Ellis et al. 2000).For many long-lived vertebrates, variation in newborn and juvenile survival probabilities may have a higher impact on population growth rate than on other vital rates (Eberhardt 1985, Crone 2001, Underwood et al. 2008).
An individual's condition is often used to understand factors influencing population dynamics, including environmental stress, susceptibility to diseases, and predation (Kuiken et al. 1994, Burek et al. 2005, Le Boeuf and Crocker 2005).Animals in better condition can invest more energy in immune defense, which may increase fitness and lead to positive population growth (Moller et al. 1998).Condition has been correlated with maternal investment (Wheatley et al. 2006) and genetic factors (Kretzmann et al. 2006) in marine mammals.By monitoring the health of a population, it is possible to determine when it may become imperiled and to identify necessary conservation strategies to ensure survival.Measuring body condition in at-risk populations may be an efficient conservation tool.For example, body condition of harbor porpoises (Phocoena phocoena) dying from disease and parasites differs from those suffering from starvation, enabling specific preventative actions to halt further changes in population health (Kuiken et al. 1994).Even for populations that are not in a state of decline, it may be important to determine a link between condition and behavior for future management.Many emerging diseases have the possibility of causing population declines (Scott 1988, Osterhaus et al. 1990); thus, determining a
La condición de un individuo es frecuentemente utilizada para entender los factores que afectan la dinámica poblacional, incluyendo el estrés ambiental, la susceptibilidad a enfermedades y la depredación (Kuiken et al. 1994, Burek et al. 2005, Le Boeuf y Crocker 2005).Los animales en mejor condición pueden canalizar más energía a su defensa inmunológica, lo cual puede incrementar su aptitud y resultar en un crecimiento poblacional positivo (Moller et al. 1998).La condición se ha correlacionado con la inversión materna (Wheatley et al. 2006) y con factores genéticos (Kretzmann et al. 2006) en mamíferos marinos.El monitoreo de la salud de una población permite determinar cuando ésta podría estar en peligro e identificar las estrategias de conservación necesarias para asegurar su supervivencia.La medición de la condición corporal de las poblaciones amenazadas puede ser una eficiente herramienta de conservación.Por ejemplo, la condición corporal de las marsopas (Phocoena phocoena) expuestas a enfermedades y parásitos difiere de las que non-invasive measurement of condition could be important to ensuring their survival.Relationships between behavior and growth exist for vertebrates such as small mammals (e.g., Krebs et al. 1973) and bighorn sheep (Ovis canadensis; Shackleton 1973).For bighorn sheep, the time spent playing relates to individual growth (Geist 1971).These findings suggest that non-invasive measurements may be used together with direct measurements to assess the relationship between condition and growth in other vertebrate species.
In this study we seek to evaluate the relationship between neonate behavior and two measurements of neonate health-body condition and growth-as a function of the population health of California sea lions, Zalophus californianus, in the Gulf of California, Mexico.Current methods to estimate body condition and growth of neonates require capture and handling.Frequent capturing activities may cause disturbances within rookeries, and frequent disturbances at rookeries could disrupt nursing behavior, as observed in harbor seals (Allen et al. 1984).Disturbances may also reduce pup survival by causing trampling or abandonment of pups (Richardson et al. 1995).If neonate behavior reflects individual condition, determining the relationship between behavior and body condition or growth may reduce stress and disturbances associated with capture activities (e.g., capture and handling), and facilitate the development of a useful tool for monitoring population health.However, handling sea lion pups to estimate growth, body condition, and additional measurements is an activity that may not be eliminated entirely because it is necessary for other research goals.We propose that using a combination of invasive and noninvasive approaches would reduce the potential for negative impacts of handling activities on individual condition.For the purposes of estimating pup growth rates and body conditions, we needed to collect pup weight, length, and girth data in the field (Young et al. 2007, see methods section).The behaviors and morphometric data we collected are typical of data available for pinnipeds, and our results may apply to species with similar life history traits.
El objetivo del presente estudio fue evaluar la relación entre el comportamiento neonatal y dos medidas de la salud neonatal-condición corporal y crecimiento-como función de la salud de la población del lobo marino de California, Zalophus californianus, en el Golfo de California, México.Los métodos actuales para estimar la condición corporal y el crecimiento de neonatos implica su captura y manipulación.Frecuentes actividades de captura pueden causar disturbios en las colonias, y tales disturbios pueden perturbar los patrones de amamantamiento, como se ha observado en focas comunes (Allen et al. 1984).Estos disturbios también pueden reducir la supervivencia de las crías al provocar el aplastamiento o abandono de las crías (Richardson et al. 1995).Si el comportamiento neonatal refleja la condición individual, entonces el determinar la relación entre el comportamiento y la condición corporal o crecimiento podría ser importante para reducir el estrés y los disturbios asociados con las actividades de captura (e.g., captura y manipulación) y facilitar el desarrollo de una herramienta eficaz para el monitoreo de la salud poblacional.Aun así, la manipulación de los lobos marinos para evaluar el crecimiento y la condición corporal, así como obtener otras mediciones de las crías, es una actividad que no se puede eliminar por completo ya que se requiere para otros objetivos de investigación.Proponemos que el uso de una combinación de métodos invasivos y no invasivos podría reducir la posibilidad de los efectos negativos asociados con las actividades de manipulación sobre la condición de los individuos.Para estimar las tasas de crecimiento y las condiciones corporales de las crías, fue necesario recolectar datos sobre su peso, longitud y ancho corporal en el campo (Young et al. 2007, ver sección de métodos).Los datos morfométricos y de comportamiento que se recolectaron son típicos de datos disponibles para pinípedos, y nuestros resultados podrían aplicarse a otras especies con características de historia de vida similares.largest of the three rookeries and comprised of rocky beaches with large boulders.Granito is centrally located and comprised of both rocky and sandy beaches backed by vertical cliffs and large boulders.Because it is the most remote of the three islands, it has a lower rate of human disturbance than Los Islotes and San Jorge (LR Gerber, unpublished data).Los Islotes is located in the southern part of the GC.Formed originally from volcanoes, the beaches are comprised predominately of rock platforms surrounded by rocky inlet channels.Mean temperatures at each island during the study ranged between 32ºC and 39ºC.

Surveys and animal measurements
Data were collected from two replicate sites on each island during two nine-day periods in June and July of 2005, 2006, and 2007.At each site we observed sea lions from a platform approximately 30-50 m away from the nearest animal to minimize disturbance (Labrada-Martagón et al. 2005).
To measure pup growth and condition during their first few months of life, individuals were captured in June and recaptured in July (table 1).Individuals were approximately one month old when they were captured in July; mean peak of births occurs ~15-21 June (García-Aguilar and Aurioles-Gamboa 2003).We recorded body length, weight, and girth for each pup that was captured.Girth was calculated by measuring the circumference around the thorax (Luque and Aurioles-Gamboa 2001).Pups were given a unique plastic flipper tag (Dalton ID Systems Jumbo Tags) for long-term identification.All capture and handling techniques were approved by the Institutional Animal Care and Use Committee at Arizona State University (permit 07-918R to Gerber).
Growth rate was calculated as (w 2w 1 )/d, where w 1 is the weight at first capture, w 2 is the weight at second capture, and d is the number of days between weighing events (Ono et al. 1987).Thus, growth rates are only available for tagged pups that we were able to capture in June and recapture in July.We used density (weight/volume) to determine the overall condition of pups sampled (Luque and Aurioles-Gamboa 2002).Volume was calculated as: (l × 0.0265) × G t 2 , where l is the length, G t is the thoracic girth, and 0.0265 is a constant derived from calculations for Weddell seals, Leptonychotes weddelli (Castellini andKooyman 1990, Castellini et al. 1993).
Behavioral observations were conducted 1-5 days before the July captures so we could accurately correlate the behaviors observed in July with current pup condition and observed growth rates that were calculated from the June and July captures.We conducted four 30-min focal observations in 2005 and four 15-min focal observations in 2006 and 2007 on four marked sea lion pups per day (Altmann 1974).We reduced the length of focal observations between years based on a preliminary power analysis that indicated 15-min intervals provided sufficient data for statistical analysis (LR Gerber, pers.comm.).For focal observations collected in 2006, we randomly selected a start time within the first en junio y recapturadas en julio.Se determinó la densidad (peso/volumen) para evaluar la condición general de las crías muestreadas (Luque y Aurioles-Gamboa 2002).El volumen se calculó de la siguiente forma: (l × 0.0265) × G t 2 , donde l es la longitud, G t es el ancho torácico y 0.0265 es una constante derivada de los cálculos realizados para la foca de Weddell, Leptonychotes weddelli (Castellini y Kooyman 1990, Castellini et al. 1993).

Análisis estadísticos
Se comprobó el supuesto de normalidad de los datos, y se utilizaron modelos lineales generalizados para evaluar el efecto del sexo, el año, la isla y el comportamiento (i.e., activo vs amamantamiento o inactivo vs amamantamiento, 15 min of each 30-min observation and used the proceeding 15-min period for all analyses.Thus, all pup focal observations used for analysis were for 15-min observations.In all years, pups were selected each day at random, without replacement, and focals were conducted at different times of the day.We used hand-held computers with observation software (Noldus Pocket Observer) to record pup behavior (table 2).This allowed behaviors to be recorded to the nearest one-hundredth second, although we rounded duration to the nearest minute.Observers were trained in classifying pup behavior prior to each field trip so they were able to identify different pup behaviors and use the hand-held computers for this study.Focal observations were removed from final analyses if the pup moved out of sight or if a behavior could not be determined.For statistical analyses, we grouped behaviors into three categories based on energy gains and losses, following Ono and Boness (1996): active (aggression, female-pup calls, fidgeting, fighting, locomotion, nursing intent, nuzzling, play, and scratching self), inactive (flipper up, resting, and sitting), and nursing.Aggregating behaviors reduced the chance of errors arising from observers classifying more fine-scale behaviors incorrectly.

Statistical analyses
We tested our data for normality and they met this assumption.We thus used generalized linear models (GLM) to explore the effect of sex, year, island, and behavior (i.e., active vs nursing or inactive vs nursing, global model) on sea lion pup condition and growth rates.Each replicate site within each individual island was aggregated because we found no statistical difference between sites at a given island.We fitted multiple linear regression models and used stepwise regression to identify the best model (Kutner et al. 2005).Dummy variables were created for categorical data (i.e., island and year).All statistical analyses were performed in SAS, and results were defined as significant if P ≤ 0.05.
El modelo global mostró un ajuste relativamente bueno a los datos para los machos (activo vs amamantamiento, R 2 = 0.5; inactivo vs amamantamiento, R 2 = 0.5), pero no para las hembras (activo vs amamantamiento, R 2 = 0.2; inactivo vs amamantamiento, R 2 = 0.2).Todas las variables fueron significativas excepto isla entre San Jorge y Los Islotes (P = 0.74) en el modelo de activo vs amamantamiento, e isla entre San Jorge y Los Islotes (P = 0.85) e inactivo (P = 0.1) en el modelo inactivo vs amamantamiento (P = 0.1) para los machos.Al eliminar las variables no significativas del modelo global, el ajuste del modelo alternativo fue muy bajo (R 2 = 0.2), por lo que se concluyó que el modelo global era el mejor.Por otro lado, todas las variables excepto isla entre San Jorge y Los Islotes fueron significativas para our analyses on pup growth.We found that sex (active vs nursing, P = 0.05; inactive vs nursing, P = 0.05) and island (active vs nursing and inactive vs nursing, P = 0.0001) had a significant effect on growth.We separated our dataset by sex to determine whether the same variables (year, island, and behavior) had an effect on growth once the effect of sex had been removed from the model.We were unable to extract the effect of sex from island because the sample size would have been significantly reduced.
The global model represented a relatively good fit to our data for males (active vs nursing, R 2 = 0.5; inactive vs nursing, R 2 = 0.5) but not for females (active vs nursing, R 2 = 0.2; inactive vs nursing, R 2 = 0.2).All variables were significant except island between San Jorge and Los Islotes (P = 0.74) in the active vs nursing model, and island between San Jorge and Los Islotes (P = 0.85) and inactive (P = 0.1) in the inactive vs nursing model (P = 0.1) for males.When nonsignificant variables were dropped from the global model, the fit of the alternative model was very low (R 2 = 0.2), so we concluded that the global model was the best.On the other hand, all variables except island between San Jorge and Los Islotes were significant for both active vs nursing (P = 0.03) and inactive vs nursing (P = 0.03) in females.Active behaviors were negatively correlated with male growth ( = -0.0005),whereas nursing was positively correlated with this parameter ( = 0.002; active vs nursing model).Inactive and nursing behaviors were both positively correlated ( = 0.0004 and  = 0.002, respectively) with male growth.amamantamiento vs activo (P = 0.03) y vs inactivo (P = 0.03) en hembras.Se observó una correlación negativa entre el comportamiento activo y el crecimiento de machos ( = -0.0005),mientras que amamantamiento estuvo positivamente correlacionado con este parámetro ( = 0.002; modelo de activo vs amamantamiento).El comportamiento inactivo y el de amamantamiento estuvieron positivamente correlacionados ( = 0.0004 y  = 0.002, respectivamente) con el crecimiento de los machos.
La variable sexo fue el factor que más afectó las tasas de crecimiento de los lobos marinos de California en el Table 2. Description of the behaviors monitored to determine the potential relationships between the behavior of California sea lion pups and their individual growth rates and body conditions.Tabla 2. Descripción de los comportamientos estudiados para determinar las relaciones potenciales entre el comportamiento de las crías del lobo marino de California y sus tasas de crecimiento y condiciones corporales.

Behavior Description Aggression
Open mouth displays and vocalizations from one individual to another without physical contact, threats.
Female-pup calls Female and pup calling for each other in search of recognition.
Fidgeting Individuals moving around and accommodating their posture, not specifically directional locomotion or resting.
Fighting Individuals grabbing, pushing, or biting one another.Physical contact must be present.

Flipper up
Animal in a resting position but with front and/or hind flippers elevated vertically.

Locomotion
Movement from one place to another (walking, running, or swimming).

Nursing
Pup or juvenile actively suckling from a female.

Nursing intent
Pups touching females with their nose trying to achieve successful nursing; no nursing occurs.
Nuzzling Individuals nudging, passing snout repeatedly over each other, sniffing each other, and scratching.Also includes females and pups touching noses as 'recognition' behavior.

Play
Pups or juveniles simulating adult behaviors together or splashing.

Resting
Animal is lying down, both front and hind flippers resting on ground.

Scratching self
Grooming or scratching self with flippers or against a substrate.

Sitting
The animal's front flippers are supporting upper body, lower body resting on ground.

DISCUSSION
Identifying non-invasive relationships between animal condition and behavior is important in light of the effects of human disturbance on animal behaviors (e.g., Allen et al. 1984, Boydston et al. 2003, Rode et al. 2006).In California sea lions, human disturbances in the form of tourists and fishermen influence reproductive success and population growth rates (French et al., in press).Thus, the development of non-invasive approaches to assess population health may assist resource managers interested in determining current population demography and health without influencing current and future population trends that may be affected by human disturbance.
The sex variable was the most important factor influencing growth rates in California sea lions in the GC.We also found that behavioral surrogates were positively related to growth rates of male pups, although there was no relationship between behaviors of females and their growth or condition.Our measurements may be more accurate in predicting male growth rates because of their faster growth within the first few months of life (Ono and Boness 1996).Female pups grow more steadily throughout their first year, whereas male pups grow at a much faster rate for a longer period of time due to differing metabolic rates, larger size at birth, and increased milk intake compared with female pups, making males more efficient at nursing (Ono and Boness 1996).Behavioral correlates of female growth may still exist, but on a year-long time scale rather than during the neonate time period measured in this study, and warrant further investigation.
Although our findings indicate that neonate growth of females is not correlated with our selected measurements of behavior, the significant relationship between measurements of male pup behaviors and their growth rates suggests that measurements of behavioral activity may serve as a proxy to measurements of individual growth rates.Thus, we have gained insight into the relationship between the behavior of male pups and growth rates that could benefit studies of sea lion population trends.Our findings suggest that resource managers interested in assessing the population health of sea lions could use non-invasive behavioral observations of male pups as an alternative to capture and handling techniques, reducing overall human disturbance.
Understanding the relationship between male behavior and growth using our non-invasive measurements may enhance our understanding of the population as a whole and improve estimates of future viability (Gerber 2006, Le Galliard et al. 2006).By acquiring this information using non-invasive measurements, we can gain insight without the potentially negative consequences of frequently using more invasive techniques.Our approach may be particularly important for pinnipeds that are more sensitive to human disturbances (e.g., seals) by providing a non-invasive method for researchers to obtain information on their current status GC.También se observó una relación positiva entre el comportamiento y las tasas de crecimiento de los neonatos machos, pero no se observó una relación entre el comportamiento de las hembras y su crecimiento o condición.Nuestras mediciones posiblemente sean más precisas para predecir las tasas de crecimiento de los machos ya que éstos crecen de forma más rápida durante sus primeros meses de vida (Ono y Boness 1996).Las crías hembras crecen de forma más constante durante su primer año de vida, mientras que los machos crecen a un paso más rápido durante un periodo de tiempo más largo debido a sus distintas tasas metabólicas, mayor tamaño al nacer y mayor consumo de leche en comparación con las hembras, lo cual los hace más eficientes para amamantar (Ono y Boness 1996).Podrían existir correlaciones entre el comportamiento y el crecimiento de las hembras, pero a escala de un año más que durante el periodo neonatal registrado en este estudio, las cuales ameritan mayor investigación.
Aunque nuestros resultados indican que el crecimiento de las hembras neonatas no se correlaciona con los comportamientos seleccionados, la fuerte correlación entre las tasas de crecimiento y el comportamiento de las crías de machos sugiere que las mediciones de su comportamiento pueden ser representativos de las mediciones de las tasas individuales de crecimiento.La información obtenida sobre la relación entre el comportamiento y las tasas de crecimiento de las crías macho podría ser útil para otros estudios de las tendencias poblacionales de lobos marinos.Los resultados sugieren que los administradores de recursos, interesados en evaluar la salud de la población, podrían realizar observaciones no invasivas del comportamiento de los neonatos machos como una alternativa de las técnicas de captura y manipulación, reduciendo de esta manera el disturbio humano.
La evaluación de la relación entre el comportamiento y el crecimiento de los machos usando técnicas no invasivas mejoraría nuestro conocimiento de la población global así como las estimaciones de su viabilidad posterior (Gerber 2006, Le Galliard et al. 2006).Si se adquiere esta información mediante mediciones no invasivas se evitarían las potenciales consecuencias negativas frecuentemente asociadas con las técnicas invasivas.Nuestro enfoque podría ser particularmente importante en el caso de los pinípedos que son más sensibles a los disturbios humanos, como las focas, al proporcionar un método no invasivo que permitiría obtener información de su estatus actual (Allen et al. 1984, Richardson et al. 1995).Sin embargo, debido a las incertidumbres en la evaluación de las tasas de crecimiento de las crías hembras, las actividades de captura y manipulación probablemente sigan siendo necesarias para obtener una idea completa de la salud de la población.Por otro lado, para obtener información próxima de la salud poblacional, las mediciones no invasivas de comportamiento ofrecen un método sencillo para determinar las tasas de crecimiento de las crías macho.(Allen et al. 1984, Richardson et al. 1995).However, because of caveats associated with understanding female pup growth rates, capture and handling activities may still be necessary to yield a complete picture of population health.But for a proximate view of population health, our non-invasive measurements of behavior offer an easy method to determine growth rates of male pups.
Our findings may differ to other studies that found strong relationships between behavior and condition and growth rates (Geist 1971, Krebs et al. 1973, Shackleton 1973) for several reasons.First, our results may differ between sexes because of the starting points in terms of health and condition for male and female pups at birth.Also, we focused on the neonate period rather than on the entire period of pup development, so the relationship between those variables may be apparent if we extended our study to the entire lactation phase of approximately one year and took multiple measurements to detect differences in those variables.Finally, we estimated body condition using a model developed for a species of seal and it may not adequately fit sea lions.
Although our results were significant for male pups and their behaviors related to individual growth rates, further studies would enrich our understanding of this relationship by addressing behavior in different ways.Each individual pup was sampled once, and due to the relatively similar birth dates of pups in a given season, age was not considered to be a significant factor (Ono et al. 1987).However, to better incorporate behavior into demographic modeling, behavior and growth rates should be determined for all age classes and sexes.While investigating the growth rates of pups may provide a snapshot of how the pup population is doing that year, by including all age classes and growth rates it may be possible to determine the overall population health for a longer period of time.Our results contribute to the general understanding of mammalian conservation, and build upon previous studies that demonstrate relationships between behavior and growth rates (e.g., Krebs et al. 1973, Shackleton 1973).These results have broad implications for developing approaches to minimize human disturbances in animal populations, while still accurately assessing population health.Non-invasive methods are particularly relevant for populations that are sensitive to human disturbances, although further research is needed in light of the threatened status of such populations.