Limits...
Beyond carbon and nitrogen: guidelines for estimating three-dimensional isotopic niche space.

Rossman S, Ostrom PH, Gordon F, Zipkin EF - Ecol Evol (2016)

Bottom Line: Yet, other stable isotopes can provide additional power to resolve questions associated with foraging, migration, dispersal and variations in resource use.We use these results to provide guidelines for sample size in conducting multivariate isotopic niche modeling.Upon inclusion of sulfur, we see that the three dolphin ecotypes are in fact segregated on the basis of salinity and find the stable isotope niche of inshore bottlenose dolphins significantly larger than coastal and offshore populations.

View Article: PubMed Central - PubMed

Affiliation: Department of Integrative BiologyMichigan State UniversityEast LansingMichigan48824; Ecology Evolutionary Biology and Behavior ProgramMichigan State UniversityEast LansingMichigan48824; Hubbs-Sea World Research Institute3830 S. Highway A1A #4-181Melbourne BeachFlorida32951.

ABSTRACT
Isotopic niche has typically been characterized through carbon and nitrogen ratios and most modeling approaches are limited to two dimensions. Yet, other stable isotopes can provide additional power to resolve questions associated with foraging, migration, dispersal and variations in resource use. The ellipse niche model was recently generalized to n-dimensions. We present an analogous methodology which incorporates variation across three stable dimensions to estimate the significant features of a population's isotopic niche space including: 1) niche volume (referred to as standard ellipsoid volume, SEV), 2) relative centroid location (CL), 3) shape and 4) area of overlap between multiple ellipsoids and 5) distance between two CLs. We conducted a simulation study showing the accuracy and precision of three dimensional niche models across a range of values. Importantly, the model correctly identifies differences in SEV and CL among populations, even with small sample sizes and in cases where the absolute values cannot precisely be recovered. We use these results to provide guidelines for sample size in conducting multivariate isotopic niche modeling. We demonstrate the utility of our approach with a case study of three bottlenose dolphin populations which appear to possess largely overlapping niches when analyzed with only carbon and nitrogen isotopes. Upon inclusion of sulfur, we see that the three dolphin ecotypes are in fact segregated on the basis of salinity and find the stable isotope niche of inshore bottlenose dolphins significantly larger than coastal and offshore populations.

No MeSH data available.


Related in: MedlinePlus

(A) Median estimated (black lines) and true (white lines) differences in the centroid location (CL) between population one and populations two through four (corresponding to number on right side of panel). Medians are calculated using 1000 simulated datasets at each sample size n = 6 through n = 100. The shaded gray areas show the 95% CI. (B) Median probability of correctly estimating that populations – two through four (number on left side of panel) had a larger SEV compared to population one for sample sizes ranging from 6–100. (C–F) Median estimated (black lines) and true (white lines) standard ellipsoid volumes (SEV) for four populations with progressively increasing isotopic niche size. Shaded gray areas show the 95% CIs.
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4834325&req=5

ece32013-fig-0001: (A) Median estimated (black lines) and true (white lines) differences in the centroid location (CL) between population one and populations two through four (corresponding to number on right side of panel). Medians are calculated using 1000 simulated datasets at each sample size n = 6 through n = 100. The shaded gray areas show the 95% CI. (B) Median probability of correctly estimating that populations – two through four (number on left side of panel) had a larger SEV compared to population one for sample sizes ranging from 6–100. (C–F) Median estimated (black lines) and true (white lines) standard ellipsoid volumes (SEV) for four populations with progressively increasing isotopic niche size. Shaded gray areas show the 95% CIs.

Mentions: Results indicate the model produces accurate estimates of CL at all sample sizes. Differences in CL between populations also are accurately detected by the model (Fig. 1A). The model had ≥95% probability in detecting the true difference in CL values for 85% of datasets with only six samples per population when the true distance in CL was only ~1.7‰ (comparison of population one to population two). At larger differences in CL and at larger sample sizes, the average probability of detecting differences between two populations is nearperfect. Absolute estimates of CL differences are nearly unbiased at even the smallest sample sizes for all populations, showing only a slight overestimation when sample sizes are small (Fig. 1A).


Beyond carbon and nitrogen: guidelines for estimating three-dimensional isotopic niche space.

Rossman S, Ostrom PH, Gordon F, Zipkin EF - Ecol Evol (2016)

(A) Median estimated (black lines) and true (white lines) differences in the centroid location (CL) between population one and populations two through four (corresponding to number on right side of panel). Medians are calculated using 1000 simulated datasets at each sample size n = 6 through n = 100. The shaded gray areas show the 95% CI. (B) Median probability of correctly estimating that populations – two through four (number on left side of panel) had a larger SEV compared to population one for sample sizes ranging from 6–100. (C–F) Median estimated (black lines) and true (white lines) standard ellipsoid volumes (SEV) for four populations with progressively increasing isotopic niche size. Shaded gray areas show the 95% CIs.
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4834325&req=5

ece32013-fig-0001: (A) Median estimated (black lines) and true (white lines) differences in the centroid location (CL) between population one and populations two through four (corresponding to number on right side of panel). Medians are calculated using 1000 simulated datasets at each sample size n = 6 through n = 100. The shaded gray areas show the 95% CI. (B) Median probability of correctly estimating that populations – two through four (number on left side of panel) had a larger SEV compared to population one for sample sizes ranging from 6–100. (C–F) Median estimated (black lines) and true (white lines) standard ellipsoid volumes (SEV) for four populations with progressively increasing isotopic niche size. Shaded gray areas show the 95% CIs.
Mentions: Results indicate the model produces accurate estimates of CL at all sample sizes. Differences in CL between populations also are accurately detected by the model (Fig. 1A). The model had ≥95% probability in detecting the true difference in CL values for 85% of datasets with only six samples per population when the true distance in CL was only ~1.7‰ (comparison of population one to population two). At larger differences in CL and at larger sample sizes, the average probability of detecting differences between two populations is nearperfect. Absolute estimates of CL differences are nearly unbiased at even the smallest sample sizes for all populations, showing only a slight overestimation when sample sizes are small (Fig. 1A).

Bottom Line: Yet, other stable isotopes can provide additional power to resolve questions associated with foraging, migration, dispersal and variations in resource use.We use these results to provide guidelines for sample size in conducting multivariate isotopic niche modeling.Upon inclusion of sulfur, we see that the three dolphin ecotypes are in fact segregated on the basis of salinity and find the stable isotope niche of inshore bottlenose dolphins significantly larger than coastal and offshore populations.

View Article: PubMed Central - PubMed

Affiliation: Department of Integrative BiologyMichigan State UniversityEast LansingMichigan48824; Ecology Evolutionary Biology and Behavior ProgramMichigan State UniversityEast LansingMichigan48824; Hubbs-Sea World Research Institute3830 S. Highway A1A #4-181Melbourne BeachFlorida32951.

ABSTRACT
Isotopic niche has typically been characterized through carbon and nitrogen ratios and most modeling approaches are limited to two dimensions. Yet, other stable isotopes can provide additional power to resolve questions associated with foraging, migration, dispersal and variations in resource use. The ellipse niche model was recently generalized to n-dimensions. We present an analogous methodology which incorporates variation across three stable dimensions to estimate the significant features of a population's isotopic niche space including: 1) niche volume (referred to as standard ellipsoid volume, SEV), 2) relative centroid location (CL), 3) shape and 4) area of overlap between multiple ellipsoids and 5) distance between two CLs. We conducted a simulation study showing the accuracy and precision of three dimensional niche models across a range of values. Importantly, the model correctly identifies differences in SEV and CL among populations, even with small sample sizes and in cases where the absolute values cannot precisely be recovered. We use these results to provide guidelines for sample size in conducting multivariate isotopic niche modeling. We demonstrate the utility of our approach with a case study of three bottlenose dolphin populations which appear to possess largely overlapping niches when analyzed with only carbon and nitrogen isotopes. Upon inclusion of sulfur, we see that the three dolphin ecotypes are in fact segregated on the basis of salinity and find the stable isotope niche of inshore bottlenose dolphins significantly larger than coastal and offshore populations.

No MeSH data available.


Related in: MedlinePlus