Reply to A.V. Smagin: IV. Surface diffusion or random noise?
| Autor: | A. F. Sabrekov, I. E. Terentieva, Mikhail Glagolev |
|---|---|
| Rok vydání: | 2017 |
| Předmět: | |
| Zdroj: | Environmental Dynamics and Global Climate Change. 8:55-65 |
| ISSN: | 2541-9307 2218-4422 |
| DOI: | 10.17816/edgcc8155-65 |
| Popis: | In this paper, we analyze experimental estimate of surface diffusion coefficient of gas in soil conducted by prof. A.V. Smagin. The experiment consisted of methane diffusion coefficient determination under different soil saturation conditions. Regression analysis was made basing on obtained dataset to determine third-power polynomial coefficients, where soil air-filled porosity is a single independent variable and methane diffusion coefficient is a single dependent variable. Free term of the polynomial was suggested to be diffusion coefficient of gas in water-saturated soil. After the regression free term was found to be several times higher than methane diffusion coefficient in water. Based on this fact A.V. Smagin made a suggestion that surface diffusion had significant impact on methane diffusive transport in soil. In this paper, we estimate the error of this approach. It was revealed that based on original experimental data of A.V. Smagin, free term of third-power polynomial was statistically insignificant (p=0.05). Moreover, its error was order of magnitude higher than its absolute value. After its exclusion from the regression polynomial, the rest of coefficients stayed statistically significant. Furthermore, statistical modelling was used to estimate the influence of measurement uncertainty of methane diffusion coefficient in soil on regression between diffusion coefficient and soil air-filled porosity. Results of statistical modelling indicated that even basing on typical error of diffusion coefficient measurements (about 10%), the error of cubical polynomial free term calculation reached more than 100%. Therefore, it can be concluded that since the regression uncertainty is considerably higher than the absolute value of diffusion coefficient, we can't make any speculations about surface diffusion because statistically it doesn't differ from zero. Besides that, in this paper we consider questions of gas transport in soil, e.g. the inclusion of porosity in mass balance equations and importance of sorption on different timescales. Finally, we discuss several problems of scientific ethics. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|