Trend Evaluation

H-GAC adopted guidance from the TCEQ regarding assessment of the statistical significance of inferred trends and accepts a regression line slope p-value of 0.10 or less as indicative of significance. It should be noted at the outset that this only establishes that there is less than a 10 percent (0.10) chance that the slope is actually zero (that is, a flat line). As the dataset grows larger, due to pooling of data from multiple stations in the watershed, trends may be identified which are neither important nor interesting. In order to minimize the bias created by large datasets and extreme values, the results of all regression analyses were compared to identify trends that were most likely to be indicative of actual changes in the watershed, rather than an effect of sample size or extreme values. In addition, possible significant trends were graphed and the distribution of the data was evaluated through examination of box plots and nonparametric analysis of variance. A total of 51 significant trends were identified in this manner and will be discussed as appropriate in the watershed water quality narrative. Graphs of those trends identified as statistically significant at several (but not all) levels of analysis may be found in the statistical appendix. All statistical analyses were performed using SAS® software. The code written for the analyses may be found in the statistical appendix.

Report Contents