JOURNAL OF PLANKTON RESEARCH | VOLUME 17 | NUMBER 12 | PAGES 2175-2190 | 1995
© Oxford University Press
research-article |
Alternative method for the calculation of mean time for the assessment of secondary production by true cohort analysis
Facultad de Ciencias Marinas, Universidad Autónoma de Baja California Apartado Postal 453, Ensenada, Baja California, México Email: atrujo{at}bahia.ens.uabc.mx
Received on April 4, 1994; accepted on June 6, 1995
The development time for the stages of a given population is a key parameter in the estimation of secondary production by the true cohort analysis method and requires an accurate knowledge of it. For this calculation, it is necessary to determine the mean time. This paper suggests and develops an alternative method for the evaluation of the mean time of the organism-time curve. To demonstrate the application and validity of this method, data on three generations of the marine calanoid copepod Calanus pacificus Brodsky were used in the Deep Tank of the Scripps Institution of Oceanography for 58 days (30 April–26 June 1981). Once calculated, the development time of the stages was assessed. A comparison of the results with the existent method of Landry (Int Rev. Ges. Hydrobiol, 63, 77–119, 1978) is carried out which mainly uses iterations with respect to time for approximating the value of the mean time. A statistical comparison was carried out with the results obtained by both methods; Landry's procedure with iterations of 0.8, 1.0, 1.5, 2.0, 4.0 and 12.0 h. Through a paired test, the differences in mean times and development times for the three generations obtained by both methods were statistically non-significant (
= 0.05) only for the Landry's time increment of 1.5 h: P = 0.432, 1.000 and 0.266 for mean time, and P = 0.820, 0.956 and 0.765 for development time, for generations 1, 2 and 3, respectively. With Landry's procedure, it is observed that while the time intervals are small or large from –4.16% of the sampling interval used, the results differ from those obtained with the method proposed here, underestimating or overestimating not only the mean time, but also all the parameters that depend on it. It is considered that the mean time obtained by the 'quadratic' method developed here is more precise and simple than the iterative one because its evaluation is based on only one interpolation, that it is really a median, is accurate.