Laurence C. Breaker

A Proposed Definition for Vector Correlation in Geophysics: Theory and Application

Abstract A universally accepted definition for vector correlation in oceanography and meteorology does not presently exist. To address this need, a generalized correlation coefficient, originally proposed by Hooper and later expanded upon by Jupp and Mardia, is explored. A short history of previous definitions is presented. Then the definition originally proposed by Hooper is presented together with supporting theory and associated properties. The most significant properties of this vector correlation coefficient are that it is a generalization of the square of the simple one-dimensional correlation coefficient, and when the vectors are independent, its asymptotic distribution is known; hence, it can be used for hypothesis testing. Because the asymptotic results hold only for large samples, and in practical situations only small samples are often available, modified sampling distributions are derived using simulation techniques for samples as small as eight. It is symmetric with respect to its arguments a...

A neural network as a nonlinear transfer function model for retrieving surface wind speeds from the special sensor microwave imager

A single, extended-range neural network (SER NN) has been developed to model the transfer function for special sensor microwave imager (SSM/I) surface wind speed retrievals. Applied to data sets used in previous SSM/I wind speed retrieval studies, this algorithm yields a bias of 0.05 m/s and an rms difference of 1.65 m/s, compared to buoy observations. The accuracy of the SER NN for clear (low moisture) and cloudy (higher moisture/light rain) conditions equals the accuracy of NNs trained separately for each of these cases. A new moisture retrieval criterion based on a single, physically interpretable parameter, cloud liquid water, is proposed in conjunction with the SER NN. Using this retrieval criterion, (1) a moisture retrieval threshold for cloud liquid water of 0.5 kg/m 2 was estimated, and (2) 40% of the data rejected by previous rain flags could be recovered. When the SER NN was trained using this retrieval criterion, a bias of 0.03 m/s and an rms value of 1.58 m/s were obtained and only 2% of the data were rejected. Also, a slight improvement in retrieval accuracy for cloudy conditions was achieved (∼10%) by including SSM/I brightness temperatures at 85 GHz. Finally, the limitations of NN algorithms are discussed in light of the present application.

Oceanic variability off the Central California coast

Abstract Mesoscale variability off the Central California coast is strongly influenced by coastal upwelling and related processes. Off Point Sur, there is significant space-time variability in oceanic properties over periods of days and distances of several tens of km. However, the internal density field, averaged over space (ca. 120km alongshore) and time (ca. 18 days), reveals the expected characteristics of a coastal upwelling regime, including an equatorward surface jet 25km to 40km offshore and weaker poleward flow below 150m and within 20km of the coast. Upper ocean circulation is strongly influenced by the bathymetry offshore to water depths of at least 1000m. Based on repeated horizontal grids of oceanographic stations, horizontal correlation scales for temperature are of the order of 30km in the upper 100m; hence, a horizontal sampling rate of ca. 10km is desirable. Vertical correlation scales for temperature are of the order of 100m; hence, a vertical sampling rate of ca. 10m is desirable. Similarly, there are indications of a temporal correlation scale of a few weeks; hence, a temporal sampling rate of ca. a few days is desirable. Empirical Orthogonal Functions (EOFs) in the cross-shore vertical plane suggest the influence of coastally-trapped motions within 40km of the coast, with an intense alongshore, near-surface, transient coastal jet located ca. 20km offshore. Sudden spring transitions to coastal upwelling conditions are often pronounced in coastal temperature records, and they may precede the offshore migration of the major upwelling front by at least a month or more. The spring transitions occur over a period of about a week. In 1980, the spring transition propagated poleward along the California coast at a speed of ca. 64km d −1 . During spring and summer, movement of a coastal upwelling front, often located between 15km and 50km offshore, is a major contributor to cross-shore variability. The upwelling front meanders with an alongshore scale of about 80km, a time scale of 30 to 40 days, and an amplitude scale of 10km RMS, based on satellite infrared (IR) imagery. The along-shore mean position of the major upwelling front migrates offshore during spring and summer. Offshore displacement of the front may be caused by Ekman transport over periods of days-to-weeks, and by Rossby wave dispersion over longer periods. During summer, a time-dependent offshore scale based on Rossby wave dynamics was more appropriate than the baroclinic Rossby radius. Every three to four years, during 1971 through 1984, El Nino episodes produced interannual variability in coastal sea surface temperature (SST) off Central California. Most of the warming occured during fall and winter. El Nino episodes were often followed by major spring transitions. The recent 1982/1983 episode was 2 to 3°C warmer than the three previous episodes over the past 14 years, and its influence lasted almost twice as long.

A Neural Network Multiparameter Algorithm for SSM/I Ocean Retrievals: Comparisons and Validations

Abstract A new empirical multiparameter Special Sensor Microwave/Imager (SSM/I) retrieval algorithm based on the neural network approach, which retrieves wind speed, columnar water vapor, and columnar liquid water simultaneously using only SSM/I brightness temperatures, is compared with existing global SSM/I retrieval algorithms. In terms of wind speed retrieval accuracy, the new algorithm systematically outperforms all algorithms considered under all weather conditions where retrievals are possible with an algorithm rms error of 1.0 m/s under clear and 1.3 m/s under clear plus cloudy conditions. It also generates high wind speeds with acceptable accuracy. This improvement in accuracy is coupled with increased areal coverage with obvious benefits for operational applications. With respect to columnar water vapor and columnar liquid water, the new algorithm reproduces the results of existing algorithms closely. The new algorithm has been tested and accepted for operational use at the National Centers for Environmental Prediction (NCEP) producing a positive impact on forecast winds through assimilation into NCEPs numerical weather prediction models.

A 40–50 day oscillation in sea-surface temperature along the central California coast

Abstract Previous studies of the tropical troposphere indicate the presence of a 40–50 day quasi-periodic oscillation in zonal winds. Spectral analysis of extensive daily sea-surface temperatures (SSTs) reveals a significant increase in variance (i.e. a quasi-periodic oscillation) between 40 and 50 days along the California coast as far south as Santa Barbara (34·3°N) and at least as far north as Port Arena (39°N). Increases in the amplitude of this oscillation at Granite Canyon (36°N) in some cases coincide with local El Nino warming episodes. The local surface winds, which also contain a quasi-periodic oscillation between 40 and 50 days, are coherent with, and lead, sea-surface temperature by approximately 10 days, consistent with local wind forcing. It is suggested that the 40–50 day oscillation in sea-surface temperature off central California may be related to the 40–50 day oscillation in the tropical troposphere through atmospheric teleconnections between the tropics and mid-latitudes.

Intraseasonal oscillations in sea surface temperature, wind stress, and sea level off the central California coast

Abstract The wavelet transform is used to conduct spectral and cross-spectral analysis of daily time series of sea surface temperature (SST), surface wind stress, and sea level off the central California coast for an 18-year period from 1974 through 1991. The spectral band of primary interest is given by intraseasonal time scales ranging from 30 to 70 days. Using the wavelet transform, we examine the evolutionary behavior of the frequently observed 40–50 day oscillation originally discovered in the tropics by Madden and Julian, and explore the relative importance of atmospheric vs oceanic forcing for a range of periods where both could be important. Wavelet power spectra of each variable reveal the event-like, nonstationary nature of the intraseasonal band. Peaks in wavelet power typically last for 3–4 months and occur, on average, approximately once every 18 months. Thus, their occurrence and/or duration off central California is somewhat reduced in comparison to their presence in the tropics. Although peaks in wind stress often coincide with peaks in SST and/or sea level, no consistent relationships between the variables was initially apparent. The spectra suggest, however, that relationships between the variables, if and where they do exist, are event-dependent and thus have time scales of the same order. Cross-wavelet spectra between wind stress and SST indicate that periods of high coherence (>0.90) occur on at least six occasions over the 18-year period of record. Phase differences tend to be positive, consistent with wind forcing. For wind stress vs sea level, the cross-wavelet spectra indicate that periods of high coherence, which tend to correlate with lags close to zero, also occur, but are less frequent. As with SST, the periods of high coherence usually coincide with events in the wavelet power spectra. The somewhat weaker relationship between wind stress and sea level may be due to an independent contribution to sea level through remote forcing by the ocean originating in the tropics. Finally, simple dynamical arguments regarding the lag relationships between the variables appear to be consistent with the cross-wavelet results.

Nonhydrostatic simulations of the regional circulation in the Monterey Bay area

[1] The regional circulation in the vicinity of Monterey Bay is complex and highly variable. We use a one-way coupled, nonhydrostatic version of the Dietrich Center for Air Sea Technology (DieCAST) ocean model to simulate the regional circulation. Seasonally varying local wind stress, topographic irregularities, coastal upwelling, and forcing from the open ocean are all important in this region. Satellite imagery often shows a cyclonic eddy inside the bay and an anticyclonic eddy outside the bay. The offshore anticyclonic eddy is also associated with a year-round anticyclonic eddy over the Monterey Submarine Canyon (MSC). The offshore eddy is better organized during winter. It is found that the California Undercurrent (200-400 m) does not enter the bay itself but is diverted offshore past the entrance of the bay, presumably to reform farther north along the coast. The main branch flows northward contributing to the deep anticyclonic eddy located approximately 50 km offshore of Monterey Bay. The simulations show that vertical motion is greater during summer than winter, as expected. During spring upwelling, the deep waters often upwell along the walls of the canyon and then spread and mix with surrounding waters. The deep circulation enhances mixing significantly due to the topography. We further investigate the regional circulation by comparing it with the cases where the deep canyon was filled gradually. Vortex stretching over the canyon just beyond the entrance to Monterey Bay and along the adjacent continental slopes contributes to cyclonic circulation at deeper levels. Vertical sections of velocity along the axis of MSC indicate horizontal and vertical patterns of flow that are generally consistent with past observations on the circulation of Monterey Bay.

Preliminary Results from Long-Term Measurements of Atmospheric Moisture in the Marine Boundary Layer in the Gulf of Mexico*

Measurements of boundary layer moisture have been acquired from Rotronic MP-100 sensors deployed on two National Data Buoy Center (NDBC) buoys in the northern Gulf of Mexico from June through November 1993. For one sensor that was retrieved approximately 8 months after deployment and a second sensor that was retrieved about 14 months after deployment, the pre- and postcalibrations agreed closely and fell within WMO specifications for accuracy. A second Rotronic sensor on one of the buoys provided the basis for a detailed comparison of the instruments and showed close agreement. A separate comparison of the Rotronic instrument with an HO-83 hygrometer at NDBC showed generally close agreement over a 1-month period, which included a number of fog events. The buoy observations of relative humidity and supporting data from the buoys were used to calculate specific humidity. Specific humidities from the buoys were compared with specific humidities computed from observations obtained from nearby ship reports, and the correlations were generally high (0.7‐ 0.9). Uncertainties in the calculated values of specific humidity were also estimated and ranged between 0.27% and 2.1% of the mean value, depending on the method used to estimate this quantity. The time series of specific humidity revealed three primary scales of variability: small scale (of the order of hours), synoptic scale (several days), and seasonal (several months). The synoptic-scale variability was clearly dominant; it was eventlike in character and occurred primarily during September, October, and November. Most of the synoptic-scale variability was due to frontal systems that dropped down into the Gulf of Mexico from the continental United States, followed by air masses that were cold and dry. One particularly intense event on 30 October 1993 was chosen for a more detailed analysis in terms of the characteristic return-flow cycles that occur in the northern Gulf of Mexico during fall and winter. Finally, cross-correlation analyses of the buoy data indicated that the prevailing weather systems generally entered the buoy domain from the south prior to September ; thereafter, they became more coherent and tended to enter the region from the north.

Coastal-ocean processes and their influence on the oil spilled off San Francisco by the M/V Puerto Rican

Abstract The oil tanker M/V Puerto Rican exploded on 31 October 1984 and later broke apart to produce a major oil spill in the coastal waters off San Francisco, California, USA. Oil from this spill initially moved to the SSW until 5 November, when it abruptly reversed direction and began moving rapidly to the north and then to the NNW during the following week. The oceanic processes that most likely contributed to the displacement of the oil spilled by the Puerto Rican are examined within the framework of a simple, empirical-hindcasting model. A large-scale flow component, wind drift, and tidal currents are included in the model. Wind drift, inferred by using a simple linear formulation, was the single most important factor in determining the over-all displacement of the oil. Residuals from the model, however, indicate that the winds alone could not fully account for the sudden and dramatic reversal in oil movement that occurred on 5 November 1984. This reversal was surge-like and coincided with an increase in sea level along the central California coast. Finally, the close agreement between the local and advective changes in sea-surface temperature in the Gulf of the Farallones at the time of the Puerto Rican oil spill indicate, although not conclusively, that this reversal could have been related to the onset of the Davidson Current or other larger-scale flow phenomena.

GOES-8 Imagery as a New Source of Data to Conduct Ocean Feature Tracking☆

Abstract Sequential imagery from the AVHRR has been used to conduct ocean feature tracking since the early 1980s. One of the primary limitations of AVHRR data for feature tracking is the lack of temporal continuity, since it is only possible to obtain coverage from the same satellite once every 12 hours. Thus, for the highly variable flows that are often encountered in coastal areas, undersampling can be a serious problem. With the availability of imagery every half hour from the new imager on the GOES-8 and -10 satellites, the possibility of tracking features on shorter time scales should be considered. Also, the higher sampling rate of the GOES imager could be particularly beneficial in obtaining cloud-free coverage of the ocean. However, unlike the AVHRR, which has 1-km resolution, the new imager on GOES has 4-km resolution in the infrared channels. Thus, even for relatively vigorous currents, it will take at least several hours for a feature to be advected over a distance of one pixel, and considerably longer to generate displacements that can be estimated reliably. Also, one of the basic assumptions in conducting ocean feature tracking has been that it is the submesoscale features that serve as the primary tracers of the flow. As pixel size increases, the ability to resolve and track features at these scales clearly comes into question. Additionally, as with AVHRR imagery, the ability to accurately navigate successive images is crucial to making reliable estimates of the feature displacements. These and other related issues are discussed, and three examples of feature tracking using imagery from the imager on GOES-8 are presented, together with qualitative verifications in each case. Finally, a new method for rapidly renavigating satellite imagery is presented.