64Minimum regression fit requirements

64 Minimum regression fit requirements

(1) This section specifies the requirements that must be met before an allometric function can be used to estimate biomass from project trees.

(2) An allometric function may be used to estimate biomass within a particular allometric domain only if the allometric function is documented in an allometric report in accordance with section 66 and if:

(a) the regression relationship upon which the allometric function is based:

(i) is statistically significant; and
(ii) achieves a coefficient of determination (r2) no less than 0.75; and

(b) the mean of the weighted residuals calculated at Equation 32b is not statistically significant from zero, as determined through applying a student t‑test with two‑tailed probability level of <0.05; and

(c) weighted residuals are normally distributed around zero.

(3) If the requirements specified in subsection (2) are not met, the project proponent may apply one of the following processes:

(a) redefine the allometric domain ex post so that the allometric function meets the requirements of subsection (2);

(b) use multiple regression techniques, with the application of multiple predictor measures so that the allometric function satisfies the requirements of subsection (2);

(c) undertake the following steps:

(i) conduct further sampling using the processes described at Subdivision 13 to Subdivision 15;
(ii) combine the data obtained from the further sampling specified in subparagraph (i) with the original dataset;
(iii) re‑perform the regression analyses specified in section 62 with reference to the combined allometric dataset specified in subparagraph (ii).

(4) In this section:

statistically significant means a two‑tailed probability level of <0.05.