Volume 17, Issue 4 (2-2024)
مرتع 2024, 17(4): 550-569 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Rostampour M. Calculation of the number of required samples to estimate Atriplex canescens (Pursh) Nutt phytomass, in an indirect method. مرتع 2024; 17 (4) :550-569
URL: http://rangelandsrm.ir/article-1-1193-en.html
URL: http://rangelandsrm.ir/article-1-1193-en.html
Department of Rangeland and Watershed Management and Research Group of Drought and Climate Change, Faculty of Natural Resources and Environment, University of Birjand, Birjand
Abstract: (280 Views)
Background and objectives: Due to the extensive area of rangelands and constraints in time and cost, direct estimation of rangeland production through plot sampling becomes impractical. Therefore, indirect methods, such as estimating phytomass from plant dimensions, have been proposed. However, determining the optimal sample size for the clipping and weighing method poses a challenge. This study aimed to determine the minimum required sample size for this method and explore its impact on correlation and regression tests.
Methodology: The study was conducted in four areas within the carbon sequestration project of Hossein Abad Sarbisheh, South Khorasan Province. Sampling was carried out during the full growth period of Atriplex canescens (Pursh) Nutt, a key plant species in the project. Initial sampling involved 10 samples to calculate the required sample size. A random individual of Atriplex was selected in each row, and its dimensions were measured. The entire aerial part of the plant was then clipped to estimate phytomass, which was subsequently weighed after drying. Correlation and regression tests were conducted to assess the relationship between plant dimensions and phytomass. Sample size was determined using power analysis with effect size, test power (60% and 80%), significant level of 0.05, and alternative hypothesis. The sample size was also determined through the drawing method. Subsequent samplings were conducted with increased sample sizes of 20 and 30. Coefficient of variation (CV) was used to evaluate sampling accuracy, and various model selection criteria were employed after regression analysis.
Results: Initial sampling of 10 samples showed no significant correlation between height and phytomass of Atriplex (R = 0.45, t (8) = 1.42, p > 0.05 (2-tailed)). However, with increased sample sizes of 20 and 30, significant correlations were observed (R = 0.67, t (18) = 3.84, p < 0.01 (2-tailed)) and (R = 0.62, t (28) = 4.15, p < 0.01 (2-tailed)), respectively. The power of correlation and regression tests increased with larger sample sizes (correlation test power: 0.27, 0.93, and 0.99; regression test power: 0.29, 0.97, and 0.98, respectively). Harvey-Collier test indicated significant linear relationship between height and phytomass in the data of 20 samples (p > 0.05), but not in 30 samples (p < 0.05). A sample size of 23 to 36 was recommended to achieve 60% and 80% power. Sampling accuracy in plant height measurement was higher than phytomass, and it decreased as the number of samples increased.
Conclusion: The power of correlation and regression tests from 10 initial samples was found to be less than 0.50%, which is generally not recommended for conducting studies. The study suggests that using 10 samples is not advisable. Additionally, increasing sample size led to higher coefficient of variation (CV) for height and phytomass, indicating reduced sampling accuracy. In rangeland research, fatigue and decreased precision in subsequent estimations and samplings were observed. The drawing method was found insufficient in determining the required sample size for estimating Atriplex phytomass.
Methodology: The study was conducted in four areas within the carbon sequestration project of Hossein Abad Sarbisheh, South Khorasan Province. Sampling was carried out during the full growth period of Atriplex canescens (Pursh) Nutt, a key plant species in the project. Initial sampling involved 10 samples to calculate the required sample size. A random individual of Atriplex was selected in each row, and its dimensions were measured. The entire aerial part of the plant was then clipped to estimate phytomass, which was subsequently weighed after drying. Correlation and regression tests were conducted to assess the relationship between plant dimensions and phytomass. Sample size was determined using power analysis with effect size, test power (60% and 80%), significant level of 0.05, and alternative hypothesis. The sample size was also determined through the drawing method. Subsequent samplings were conducted with increased sample sizes of 20 and 30. Coefficient of variation (CV) was used to evaluate sampling accuracy, and various model selection criteria were employed after regression analysis.
Results: Initial sampling of 10 samples showed no significant correlation between height and phytomass of Atriplex (R = 0.45, t (8) = 1.42, p > 0.05 (2-tailed)). However, with increased sample sizes of 20 and 30, significant correlations were observed (R = 0.67, t (18) = 3.84, p < 0.01 (2-tailed)) and (R = 0.62, t (28) = 4.15, p < 0.01 (2-tailed)), respectively. The power of correlation and regression tests increased with larger sample sizes (correlation test power: 0.27, 0.93, and 0.99; regression test power: 0.29, 0.97, and 0.98, respectively). Harvey-Collier test indicated significant linear relationship between height and phytomass in the data of 20 samples (p > 0.05), but not in 30 samples (p < 0.05). A sample size of 23 to 36 was recommended to achieve 60% and 80% power. Sampling accuracy in plant height measurement was higher than phytomass, and it decreased as the number of samples increased.
Conclusion: The power of correlation and regression tests from 10 initial samples was found to be less than 0.50%, which is generally not recommended for conducting studies. The study suggests that using 10 samples is not advisable. Additionally, increasing sample size led to higher coefficient of variation (CV) for height and phytomass, indicating reduced sampling accuracy. In rangeland research, fatigue and decreased precision in subsequent estimations and samplings were observed. The drawing method was found insufficient in determining the required sample size for estimating Atriplex phytomass.
Type of Study: Applicable |
Subject:
Special
Received: 2023/03/26 | Accepted: 2024/01/30 | Published: 2024/02/29
Received: 2023/03/26 | Accepted: 2024/01/30 | Published: 2024/02/29
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
