Methods for Analysis of Spatial and Temporal Patterns

Table 1. RMP and pilot study sampling stations. Table 2. RMP and pilot sampling events. Table 3. Trace element data available from sampling events 7–15 for model-based clustering of sampling sites. Table 4. Definition of a stratification scheme for the MIDAS transects in San Francisco Bay. Table 5. Sample SpaceStat output from an OLS ANOVA for ln-transformed chromium (event 13). Table 6. Summary of spatial ANOVA diagnostics. Table 7. Sample SpaceStat output from a spatial lag ANOVA for lead (event 13). Table 8. Summary of spatial ANOVA results. Table 9. Stations identified from Moran scatterplots as most important positive anomaly for each trace element during sampling event 13. Table 10. Mantel tests of association between trace element totals, TSS and spatial position for sampling event 12. TE, COG and SPACE refer to the respective distance matrices for the trace element, TSS and spatial position. X·Y denotes the Mantel statistic for X and Y. (X·Y)·Z denotes the partial Mantel statistic for X and Y given Z. Statistics significant at the p=0.05 level are in bold. Table 11. Possible models of causal relationships among space, a cognate and a trace element, assuming that the trace element does not determine cognate distribution. Table 12. Value of Kendall's tau for Mann-Kendall tests of trend during the period 1991–1995. Values in bold are significant (p<0.05). Table 12. " Spatial Kendall test " statistics resulting from treating stations as seasons and sampling events as years. Stations were subdivided into four subgroups as suggested by the individual trend tests for copper (Figure 7). Table 13. Significance (p-value) of slope for linear regression of trace element total on Net Delta Outflow. Table 14. Kendall tau statistics for test of trend in total cadmium corrected for Net Delta Outflow. Table 15. Number of sampling events per calendar year quarter. Figure 4.Total trace element distributions mapped in UTM coordinates. Each map corresponds to a single element and single RMP sampling event. The area of each square is proportional to the concentration, expressed as a fraction of the largest value found in each sampling event. Figure 5.Moran scatterplots for each trace element during sampling event 13. The dashed line is the linear regression line. The solid line is the least trimmed squares regression line. Three points are designated by their station codes rather than by circles. These have the three most negative residuals with respect to the least trimmed squares line. …