Martin Gagnon

Mapping the Greenscape and Environmental Equity in Montreal: An Application of Remote Sensing and GIS

Vegetation and green spaces provide multiple benefits for urban life but they are not always evenly distributed throughout cities. Recent studies have shown that deprived and ethnic populations have less access to vegetation, which is a form of environmental inequity. The goal of this study is to map the vegetation cover and to spatially depict the problematic areas in terms of environmental inequity in Montreal. We carry out an object-oriented classification in eCognition from Quickbird images (at a resolution of 60cm) to identify two categories of vegetation: trees/shrub and grass. We then compute 12 vegetation indicators representing the proportion of vegetation, trees/shrub and grass in streets, alleys and residential yards. Finally, statistical analyses are undertaken to reveal the link between the vegetation indicators and the proportion of immigrants, visible minorities and low income individuals. Our results show that the proportion of vegetation varies significantly across the boroughs. About 30% of the areas exhibiting an elevated proportion of the three groups are identified as very high inequity whereas 10 to 14% are identified as areas with high green benefits. Environmental inequity with respect to the three populations also expresses differently depending on the type of green spaces (street and alleys for immigrants and visible minorities while all the three types of green spaces for the low income population). This study may interest city planners and local governments as the findings could better inform decisions regarding greening programs.

Processing differences between memory search and foveal visual search

Recently, Chiang and Atkinson (1976) investigated individual differences and the interrelationships in performance between memory search and visual search. Both tasks involve encoding and comparison processes: it is likely that in memory search the slope of the reaction time (RT) vs. list length function reflects at least the comparison processes, while in visual search the slope may be related to both encoding and comparison processes. Comparing the observed slopes across individual subjects, Chiang and Atkinson found no differences between the slope for memory search (MSLOPE) and that for visual search (VSLOPE); the same was true for the intercepts of the RT functions in the two tasks. Moreover, MSLOPE and VSLOPE were significantly correlated and the regression coefficient of their relation was close to one. Given the high degree of similarity between the results in the two tasks, Chiang and Atkinson concluded (p. 668, paragraph 3) that it was possible that the same comparison process subserves the recognition response in both. In the present note, we report a similar study in which the two tasks were compared across different types of stimuli. Our results indicate that the similarities seen between the two tasks may be limited to certain experimental conditions and that, under other conditions, significant processing differences are apparent. As an index of the similarities of the processes reflected by the slopes of the two tasks, we analyzed the relationship between their slopes for both positive and negative responses (averaged over subjects) across three types of stimuli: numbers, letters, and symbols. If the same comparison process is involved in both tasks, we would expect, at first glance, that the regression coefficient between the two slope measures would be close to one. If, in the visual search task, the slope also reflects encoding processes and the encoding processes are influenced by stimulus type in the same manner as the comparison