State-space statistical models are applied to long environmental time series of monthly northward wind stress, sea surface temperature (SST), salinity (SSS), and sea level (SL) from the west coast of North America. The models use a combination of Kalman filtering and maximum likelihood methods, which estimate a nonlinear trend, a nonstationary and nondeterministic seasonal signal, and an autoregressive term, and effectively separate the seasonal signals from the long-term trends. The seasonal series are examined for behavior consistent with increasing coastal upwelling during the spring-summer upwelling season, presumably in response to a pattern of long-term global warming. Over a region of the California Current System (CCS) where coastal upwelling is a dominant process (32--40 ¿N), wind stress, SST, SSS, and SL all show strong evidence of a systematic intensification of upwelling during April--July. Model trend series suggest a linear tendency for increasing equatorward stress (in agreement with the seasonal tendency), but warmer SST (opposite the seasonal and the expectation of greater upwelling). The linear tendencies of the SST and stress trends are generally an order of magnitude greater than the seasonal tendencies. Thus the long-term trend in SST masks the cooling effect of increased seasonal upwelling, and the trend in equatorward stress suggests an artificially large seasonal increase in the observed spring and summer stress. A key to identifying these patterns has been the ability to separate the long-term nonlinear trend, using the state-space models, which mask the signal of increased upwelling in the observations.¿ 1997 American Geophysical Union