Advanced

Sample sizes for estimating key ecosystem characteristics in a tropical terra firme rainforest

Metcalfe, Dan LU ; Meir, Patrick; Aragao, Luiz Eduardo O. C.; da Costa, Antonio; Almeida, Samuel; Braga, Alan; Goncalves, Paulo; Athaydes, Joao; Malhi, Yadvinder and Williams, Mathew (2008) In Forest Ecology and Management 255(3-4). p.558-566
Abstract
This study evaluated the sample sizes necessary to estimate several soil and vegetation characteristics within 10% confidence intervals with 95% probability in three terra firme tropical rainforest sites. Across all three plots, the most spatially heterogeneous variables were measurements of total standing crop root mass, ground surface litter mass, litter fall, root growth and soil respiration which required, on average, 152, 105, 52, 45 and 28 samples, respectively to estimate mean values within 10% confidence intervals with 95% probability. Leaf area index measurements integrated canopy characteristics over a relatively large spatial area and therefore only required five samples, on average, to achieve the same degree of precision.... (More)
This study evaluated the sample sizes necessary to estimate several soil and vegetation characteristics within 10% confidence intervals with 95% probability in three terra firme tropical rainforest sites. Across all three plots, the most spatially heterogeneous variables were measurements of total standing crop root mass, ground surface litter mass, litter fall, root growth and soil respiration which required, on average, 152, 105, 52, 45 and 28 samples, respectively to estimate mean values within 10% confidence intervals with 95% probability. Leaf area index measurements integrated canopy characteristics over a relatively large spatial area and therefore only required five samples, on average, to achieve the same degree of precision. Measurements of soil temperature, moisture, carbon and nitrogen content in the surface 30 cm soil layer displayed the lowest degree of spatial variation: requiring a maximum of seven samples to estimate mean values within 10% confidence intervals with 95% probability. This study, together with a review of data from similar ecosystems, suggests that standing crop root mass, root growth, litter fall and ground surface litter mass are usually acutely under-sampled, which could impede detection and interpretation of patterns and processes in these potentially important ecosystem characteristics. This information may assist researchers to design effective sampling strategies for field experiments, particularly in tropical forests. (c) 2007 Elsevier B.V. All rights reserved. (Less)
Please use this url to cite or link to this publication:
author
publishing date
type
Contribution to journal
publication status
published
subject
keywords
amazon tropical rain forest, spatial heterogeneity, coefficient of, variation, sample size, plant biomass, root production, soil respiration
in
Forest Ecology and Management
volume
255
issue
3-4
pages
558 - 566
publisher
Elsevier
external identifiers
  • wos:000254072700019
  • scopus:39049125167
ISSN
1872-7042
DOI
10.1016/j.foreco.2007.09.026
language
English
LU publication?
no
id
d8f528ff-e391-4b9b-91f1-2192d97773a0 (old id 4644079)
date added to LUP
2014-09-23 15:23:08
date last changed
2017-08-20 04:09:47
@article{d8f528ff-e391-4b9b-91f1-2192d97773a0,
  abstract     = {This study evaluated the sample sizes necessary to estimate several soil and vegetation characteristics within 10% confidence intervals with 95% probability in three terra firme tropical rainforest sites. Across all three plots, the most spatially heterogeneous variables were measurements of total standing crop root mass, ground surface litter mass, litter fall, root growth and soil respiration which required, on average, 152, 105, 52, 45 and 28 samples, respectively to estimate mean values within 10% confidence intervals with 95% probability. Leaf area index measurements integrated canopy characteristics over a relatively large spatial area and therefore only required five samples, on average, to achieve the same degree of precision. Measurements of soil temperature, moisture, carbon and nitrogen content in the surface 30 cm soil layer displayed the lowest degree of spatial variation: requiring a maximum of seven samples to estimate mean values within 10% confidence intervals with 95% probability. This study, together with a review of data from similar ecosystems, suggests that standing crop root mass, root growth, litter fall and ground surface litter mass are usually acutely under-sampled, which could impede detection and interpretation of patterns and processes in these potentially important ecosystem characteristics. This information may assist researchers to design effective sampling strategies for field experiments, particularly in tropical forests. (c) 2007 Elsevier B.V. All rights reserved.},
  author       = {Metcalfe, Dan and Meir, Patrick and Aragao, Luiz Eduardo O. C. and da Costa, Antonio and Almeida, Samuel and Braga, Alan and Goncalves, Paulo and Athaydes, Joao and Malhi, Yadvinder and Williams, Mathew},
  issn         = {1872-7042},
  keyword      = {amazon tropical rain forest,spatial heterogeneity,coefficient of,variation,sample size,plant biomass,root production,soil respiration},
  language     = {eng},
  number       = {3-4},
  pages        = {558--566},
  publisher    = {Elsevier},
  series       = {Forest Ecology and Management},
  title        = {Sample sizes for estimating key ecosystem characteristics in a tropical terra firme rainforest},
  url          = {http://dx.doi.org/10.1016/j.foreco.2007.09.026},
  volume       = {255},
  year         = {2008},
}