erage of the pace of ENSO activity, and its connections with higher latitude. Fixed arrays, which have revealed so much about the Pacific, could be extended to Indian and Atlantic oceans, at least in skeletal form.

The ocean near the coasts is a particularly sensitive and biologically productive region. Considering the close proximity of so many humans, it is surprising how poor the instrumental record is. As natural variability is overlaid on human-induced changes of many kinds, there is a lack of baseline observations for comparison. Our definition of climate and its impacts is expanding to include ecosystems and human activity, and in this spirit there are many more observations to be made, particularly aiming at biological fields. Regular coastal observational programs are hampered by overlapping jurisdictions and variable commitment to basic scientific questions. River systems are intimately connected with estuaries, and human modification again compounds natural climate variability of streamflow, water quality, and drainage patterns. Support for sustained observations in both coastal regions and watersheds should be of high priority.