Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Meteorological influences on sea surface temperatures in Queen Charlotte Sound Faucher, Manon 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-0452.pdf [ 14.54MB ]
Metadata
JSON: 831-1.0099199.json
JSON-LD: 831-1.0099199-ld.json
RDF/XML (Pretty): 831-1.0099199-rdf.xml
RDF/JSON: 831-1.0099199-rdf.json
Turtle: 831-1.0099199-turtle.txt
N-Triples: 831-1.0099199-rdf-ntriples.txt
Original Record: 831-1.0099199-source.json
Full Text
831-1.0099199-fulltext.txt
Citation
831-1.0099199.ris

Full Text

METEOROLOGICAL INFLUENCES ON SEA SURFACE TEMPERATURES IN QUEEN CHARLOTTE SOUND by Manon Faucher B.Sc, The University of Quebec in Montreal, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF EARTH AND OCEAN SCIENCES We accept this thesis as comforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 1996 © Manon Faucher, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of H f t ^ . T U frvfr> n r . ^ * n ^ A F ^ , ^ The University of British Columbia Vancouver, Canada Date C ^ ^ , ^ DE-6 (2/88) Abstract Short-term variations of sea surface temperatures (SST) over Queen Charlotte Sound have been poorly understood, mainly due to the lack of data, and therefore hardly predictable. A potentially important consequence of SST variations is the choice of salmon homeward migratory route, which has a significant impact on commercial fisheries. Unt i l recently, predictions of fish migration routes have been made by using SST data at Kains Island, one of the lighthouse stations at the northern end of Vancouver Island. Since the early nineties, A E S buoy stations have provided a new set of hourly SST in offshore waters, which may be a better representation of the fish marine environment. This thesis is using visual inspection, statistical analysis and A V H R R satellite imagery to show that the SST at Kains Island do not represent those over the main portion of the Queen Charlotte Sound, but only the SST within 20 km to 30 km from the coast. The SST at the buoy 46207 gives a better representation of the area. Furthermore, the most significant SST variations are caused by upwell ing associated wi th an offshore high pressure system and a lee trough along the west coast of Vancouver Island. i i Table of Contents Abstract i i Table of Contents i i i Lis t of Tables v i L is t of Figures i x A c k n o w l e d g e m e n t x v Chapter 1 Introduction: 1 1.1 Mot ivat ion for this Work 1 1.2 Objectives 3 1.3 Plan of the Thesis 4 Chapter 2 Description of the Study Area: 7 2.1 B a c k g r o u n d 7 2.1.1 B a t h y m e t r y 7 2.1.2 Tides 8 2.1.3 Surface Sal in i ty 9 2.1.4 Surface Water Temperature 1 0 2.1.5 Subsurface Temperature and Salinity 11 2.1.6 Cur ren ts 1 2 2.1.7 Meteorological Forc ing 13 Chapter 3 Descript ion of Data: 3 1 3.1 Data Sources: Meteorological and Oceanographic Data 3 1 3.1.1 Lighthouse Stations 3 1 3.1.2 Buoy Data 3 2 3.1.3 Other Data 3 5 3.2 Data Descript ions 3 5 3.2.1 Temperature 3 5 3.2.1.1 Variat ions in T ime 3 5 3.2.1.2 Variations in Space 3 8 3.2.2 Synoptic and Wind Analysis 3 8 3.2.2.1 Variat ions in T ime 3 9 i i i 3.2.2.2 Variations in Space 4 6 Chapter 4 Statistical Analysis: 6 8 4.1 Regression Analysis 6 8 4.1.1 Results From Regression Analysis 6 9 4.1.2 SST Anomalies Time Series Analys is 7 2 4 .2 Spectral Ana lys is 7 3 4.2.1 Results from Spectral Analysis 7 3 4 .3 Corre lat ion Ana lys is 7 4 4.3.1 Correlat ion Coeff ic ient 7 4 4.3.2 Results from Correlation Analys is - SST Anomalies 7 5 4.3.3 Results from Correlat ion Analys is - Alongshore W i n d Component 7 7 4 .4 General Observations 7 7 Chapter 5 Background Theory: 9 5 5.1 Ver t i ca l Strat i f icat ion 9 5 5.2 Heat Budget 9 8 5.2.1 Solar Radiation 9 9 5.2.2 Long-Wave Radiat ion 1 0 0 5.2.3 Conduc t ion 101 5.2.4 E v a p o r a t i o n / C o n d e n s a t i o n 1 0 2 5.2.5 Advection 1 0 4 5.2.6 Total Heat Gain/Loss 1 0 4 5.3 Ekman Layer Theory 105 5.3.1 Upwe l l i ng Theory 1 1 0 5.3.2 Two- layer M o d e l 113 5.3.2.1 Coastal upwel l ing using the two- layer model 1 1 7 Chapter 6 Compar ison Between SST , Synoptic Weather Maps and W i n d Stress 1 2 6 6.1 Event 1: July 9th to August 8th of 1990 127 6.1.1 Overview of Synoptic Weather Situation 127 6.1.2 Compar ison Between Synoptic Pressure Pattern, Wind , Cloud and SST Data 128 6.1.3 Summary 1 34 6.2 Event 2: July 3rd to August 4th of 1992 , 1 3 6 6.2.1 Overview of Synoptic Weather Situation 1 3 6 i v 6.2.2 Compar ison Between Synoptic Pressure Pattern, Wind , Cloud and SST Data 1 3 6 6.2.3 Summary 1 4 2 6.3 Event 3: June 29th to July 29th of 1993 1 4 4 6.3.1 Overv iew of Synoptic Weather Situation 1 4 4 6.3.2 Compar ison Between Synoptic Pressure Pattern, Wind, Cloud and SST Data 1 4 4 6.3.3 Surnmary '.. 1 4 9 Chapter 7 Satellite Remote Sensing of Sea Surface Temperature 1 5 9 7.1 Data Source 1 5 9 7.2 Characteristics of Infrared Radiation 161 7.3 Satel l i te Measurements ...16 3 7.4 Upper layer thermal structure 1 6 5 7.5 Limitat ions of IR sensing of the Ocean Surface 1 6 7 7.5.1 Characteristics of the skin layer 1 6 7 7.6 Image Processing 16 8 7.7 A V H R R Observations 1 7 0 7.7.1 Image 1 171 7.7.2 Image 2 1 7 5 7.7.3 Image 3 1 7 9 7.7.4 General Observations in Upwe l l ing Condit ions... 1 83 Chapter 8 Conclusion... . . . . 1 9 6 8.1 Summary and Conclusion 196 8.2 Future Work 1 9 8 B i b l i o g r a p h y 1 9 9 Append ix A Decoding of Synoptic Report 2 0 2 Appendix B Features on Synoptic Weather Maps 2 0 5 Appendix C Synoptic Weather Maps For Event 1.... 2 0 6 Appendix D Synoptic Weather Maps For Event 2 2 3 8 . Appendix E Synoptic Weather Maps For Event 3 27 2 v List of Tables 1 Geographic locations (latitude and longitude) for the lighthouse stations 31 2 SST missing data for the lighthouse stations including the missing period and the corresponding number of days 32 3 Geographic locations (latitude and longitude) and water depth (m) for the buoy stations 32 4 SST missing data for the buoy stations including the missing period and the corresponding number of days 34 5 Statistics for SST at the lighthouse and buoy stations including the summer m in imum, maximum, mean and the standard deviation (°C) 37 6 Statistics for w ind data including the summer mean w ind speed (m / s) and direction (degree), the summer maximum wind speed (m/s) wi th the corresponding direction (degree) and the day when this maximum was recorded. The periods of missing data are also indicated .: ; 45 7 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for Kains Island 70 8 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for Mclnnes Island 70 9 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for Egg Island 70 10 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for the buoy 46207 70 11 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for the buoy 46208 71 12 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for the buoy 46204 71 13 The coefficients, amplitude and phase of the annual cycle from the regression analysis wi th the sum of residuals for the buoy 46185 ..71 14 The correlation coefficient r for the summer SST time series between the lighthouse stations. The star sign (*) indicates a coefficient r significantly different from zero 76 15 The correlation coefficient r for the summer SST time series between the buoy stations. The star sign (*) indicates a coefficient r significantly different from zero '. 76 16 The correlation coefficient r for the summer SST time series between Kains Island and the buoy stations. The star sign (*) indicates a coefficient r significantly different from zero 76 17 The correlation coefficient r for the summer SST time series between Mclnnes Island and the buoy stations 76 18 The correlation coefficient r for the summer SST time series between Egg Island and the buoy stations 76 19 The correlation coefficient r for the summer alongshore w ind time series at the buoy stations 77 20 Statistics for the w ind stress including the summer maximum, min imum, mean and the standard deviation. Values are in N / m 2 107 21 Mean, max and min SST with the standard deviation (°C) for event 1 135 22 Mean, max and min SST with the standard deviation (°C) for event 2 143 23 Mean, max and min SST wi th the standard deviation (°C) for event 3 149 24 Dai ly in situ SST and SST anomalies (°C) , and wind stress ( N / m 2 ) at Kains Island and the buoy stations for day 92 (July 31st) of 1992 172 v i i 25 Statistics for SST along the transects taken from the image 1 (day 92/July 31st of 1992) including the maximum, min imum, mean and the standard deviation (°C) 174 26 Dai ly in situ SST and SST anomalies (°C) , and w ind stress ( N / m 2 ) at Kains Island and the buoy stations for day 68 (July 7th) of 1993 175 27 Statistics for SST along the transects taken from the image 2 (day 68/July 7th of 1993) including the maximum, min imum, mean and the standard deviation (°C) 178 28 Dai ly in situ SST and SST anomalies (°C) , and wind stress ( N / m 2 ) at Kains Island and the buoy stations for day 69 (July 8th) of 1993 180 29 Statistics for SST along the transects taken from the image 3 (day 69/July 8th of 1993) including the maximum, min imum, mean and the standard deviation (°C) ; 182 v i i i List of Figures 1 a) Proport ion of adult sockeye salmon returning to the Fraser River via the northern route (through Johnstone Strait) (data from IPSFC). b) Migratory routes of adult sockeye salmon around Vancouver Island (From Groot and Quinn, 1987) 6 2 The northwest coast of Brit ish Columbia including the study area (From Crawford et al, 1995) ;. 18 3 Bathymetry of the northwest coast of B.C. including the study area. Depth are in metres. (From C. Hannah, 1992) 19 4 Co-range and co-phase values for the semi-diurnal tide. Tidal range (broken line) in metre, t idal phase (solid line) in degrees. Difference of 29° corresponds to time difference of 1 h. (From Thomson, 1981) 20 5 Salinity distribution (ppt) at 3 m depth a) for May 3rd-28th, 1954, b) for June 29th-July 22nd, 1954 and c) for August 17th-September 9th, 1954. The arrows indicate the direction of flows. (From Dodimead, 1980) 21 6 Temperature distribution in degree Celcius (°C.) at 3 m depth a) for May 3rd-June 20th, 1954, b) for June 29th-July 22nd, 1954 and c) for August 17th-September 9th, 1954. The arrows indicate the direction of flows. (From Dodimead, 1980) 22 7 Temperature, and salinity structures for waters in Queen Charlotte Sound at station A . (Circled numbers indicate years of survey). (From Dodimead, 1980) 23 8 Temperature and salinity structures for waters in Queen Charlotte Sound at station C. (Circled numbers indicate years of survey). (From Dodimead, 1980) 24 9 Prevai l ing surface currents in the North Pacific (From Thomson, 1981) 25 ix 10 July-August average near surface currents over the Queen Charlotte Sound and southern Hecate Strait. (From Crawford et al, 1995) 25 11 Mean sea level pressure pattern for January (pressure values are in mil l ibars, mb). The arrows indicate winter storm tracks. (From Env. Canada, 1991) 26 12 Mean sea level pressure (mb) pattern associated wi th a deep winter low over the Gul f of Alaska wi th the related frontal system approaching the B.C. coast. (From Env. Canada, 1991) 27 13 Mean sea level pressure (mb) pattern for July. The arrows indicate summer storm tracks. (From Env. Canada, 1991)..... 28 14 Typical summer sea-level pressure (mb) pattern, a) Summer pressure trough over southern B.C. interior, b) lee trough and c) summer front (arrows and w ind barbs indicate direction of surface winds. Wind barbs also indicate strength of winds). (From Env. Canada, 1991) 29-30 15 Sea surface temperatures in degree Celsius (°C) at lighthouse and buoy stations for the summers (day 1 (May 1st) to day 153 (September 30th)) of 1990 to 1994; 47 to 49 16 Distribution of seasonal mean SST (°C) for summers 1990 to 1994. Isotherms are drawn every 0.5°C. Missing data are indicated by "m". Relatively warm waters are indicated by "w" 50-51 17 Wind time series in metre per second (m/s) at the buoy stations a) for the summer 1990 (Day 1 (May 1st) to day 153 (September 30th)), b) the summer of 1991, c) the summer of 1992, d) the summer of 1993 and e) the summer of 1994 52 to 56 18 Synoptic weather maps for northeastern Pacif ic/western Canada wi th sea-level pressures in mill ibar, a) July 8th, 1990, b) July 14th, 1990, c) August 2nd, 1990, d) August 23rd, 1990, e) May 17th, 1991, f) May 27th, 1991, g) August 12th, 1991, h) August 24th, 1991, i) May 10th, 1992, j) May 14th, 1992, k) July 1st, 1992,1) July x 17th, 1992, m) July 24th, 1992, n) August 12th, 1992, o) August 18th, 1992, p) June 4th, 1993, q) June 7th, 1993, r) June 9th, 1993, s) July 5th, 1993, t) July 9th, 1993 57 to 67 19 SST (-) superimposed on the annual cycle of SST (--) for a) Kains Island and b) the buoy 46207. Values are in degree Ce lc ius" (°C) 79-80 20 Distribution of the amplitude of the annual cycle of SST (°C) for summers 1990 to 1994 (from regression analysis). Isotherms are drawn every 0.5°C. Miss ing data are indicated by "m". Relatively warm waters are indicated by "w" 81-82 21 Sea surface temperature a n o m a l i e s ( ° C ) at lighthouse and buoy stations for a) summer 1990 (Day 1 (May 1st) to day 153 (September 30th)). b) the summer 1991, c) the summer 1992, d) the summer 1993 and e) the summer 1994 83 to 91 22 Power spectrum density for SST anomalies at a) Kains Island, b) the buoy 46207 and c) the buoy 46204. The area within the dashed lines represents the 95% confidence intervals. The bandwidth is 0.12 cpd and the number of degrees of freedom is 6 92 to 94 23 Distribution of the stratification parameter calculated for the spring tide (left) and the neap tide (right) (Jardine et al, 1993) 121 24 Dai ly short-wave radiation Qs in watts per square metre (w/m^) received at the surface of the ocean in absence of cloud. From Pickard and Emery (1982) 122 25 Long-wave radiation Qb in watts per square metre (w/m^) from a water surface as a function of surface temperature and the overlying relative humidi ty in absence of cloud. From Pickard and Emery (1982) 122 26 Alongshore w ind stress component in Newton per square metre ( N / m^) at the buoy stations for the a) summer 1990 (Day 1 (May 1st) to day 153 (September 30th)), b) summer 1991, c) summer 1992, d) summer 1993 and e) summer 1994 123-124 x i 27 Representation of two superposed shallow homogeneous layers of f luid. H i , H 2 are the depths of the layers for a f luid at rest and H = H i + H 2 is the total depth. The z coordinate increases upward where z = ri(x,y,t) is the surface elevation and z = -Hx +h(x,y,t) is the position of the interface between the layers. From G i l l (1982) 125 28 The solution for local upwel l ing at an eastern boundary. There is a coastal jet created in the upper layer in the same direction of the w ind and an undercurrent in the opposite direction. From G i l l (1982) 125 29 W ind stress time series for event 1: July 9th-August 8th (day 70-100) of 1990. W ind stress values are in Newton per square metre (N /m^) 150 30 SST time series for event 1: July 9th-August 8th (day 70-100) of 1990. SST values are in degree Celcius (°C) 151 31 C loud cover time series for event 1: July 9th-August 8th (day 70-100) of 1990. C loud cover values are in oktas 152 32 W ind stress time series for event 2: July 3rd-August 4th (day 64-96) of 1992. W ind stress values are in Newton per square metre (N /m^) 153 33 SST time series for event 2: July 3rd-August 4th (day 64-96) of 1992. SST values are in degree Celcius (°C) 154 34 C loud cover time series for event 2: July 3rd-August 4th (day 64-96) of 1992. C loud cover values are in oktas 155 35 W ind stress time series for event 3: June 29th-July 29th (day 60-90) of 1993. W ind stress values are in Newton per square metre (N/m^) 156 36 SST time series for event 3: June 29th-Jufy 29th (day 60-90) of 1993. SST values are in degree Celcius (°C) 157 37 C loud cover time series for event 3: June 29th-July 29th (day 60-90) of 1993. C loud cover values are in oktas 158 38 Emission spectra at three different temperatures. From Robinson (1985) 184 x i i 39 Energy-wavelength spectrum of solar radiation at the surface of the ocean and at different depths (nm=nanometre=10~9m). (From Brown et al, 1989) 184 40 Typical mean temperature profile for mid-latitudes in the open oceans 185 41 Temperature profiles showing successively the growth (solid lines) and decay (broken lines) of a seasonal thermocline in the Northern hemisphere; 185 42 Typical near-surface temperature profiles showing the diurnal thermocline dur ing calm, sl ightly-mixed and night-time conditions. (From Robinson, 1985) 186 43 A V H R R satellite image for day 92 (July 31st) 1992, 22:19 PDT 187 44 Synoptic weather maps for northeastern Pacif ic/western Canada wi th sea-level pressures in mill ibar for day 92 (July 31st), 1992 at 01:00 PDT 188 45 Satellite sensed sea surface temperature (°C) distribution for day 92 (July 31st), 1992 at 22:19 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km) 189 46 A V H R R satellite image for day 68 (July 7th) 1993, 23:57 PDT 190 47 Synoptic weather maps for northeastern Pacif ic/western Canada wi th sea-level pressures in mill ibar for day 68 Quly 7th), 1993 at 01:00 PDT 191 48 Satellite sensed sea surface temperature (°C) distribution for day 68 (July 7th), 1993 at 23:57 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) 2.5 km south of Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km) 192 49 A V H R R satellite image for day 69 (July 8th) 1993, 23:45 PDT 193 50 Synoptic weather maps for northeastern Pacif ic/western Canada wi th sea-level pressures in mill ibar for day 69 Quly 8th), 1993 at 01:00 PDT 194 x i i i 51 Satellite sensed sea surface temperature (°C) distribution for day 69 (July 8th), 1993 at 23:45 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) 2.5 k m south of Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km) 195 x i v Acknowledgment I would l ike to thank my supervisor Dr . Paul LeB lond for his support and guidance throughout this project. Paul's posit ive attitude always kept me enthusiastic about my thesis work. I am also grateful to Howard Freeland for his helpful comments and for giv ing me access to part of the data. A big thank goes to Ian Jardine for his very valuable help, advice and patience al l along. I am also grateful to Denis Laplante, Danie l Ricard and the graduate students who were sharing the office space with me for their support, especial ly when computer problems arose. A special thanks to Renee Jacques and Mar ie-Claude Bourque for sharing their knowledge. Thanks to al l members of the Biophysical Controls of Salmon Migrat ion and Production project. In particular, I am grateful to Ke i th Thomson for his advice. I wish to thank my parents very profoundly for al l the support and confidence they gave me. I am very grateful to the Atmospheric Environment Service for funding this project. I acknowledge the Paci f ic & Yukon Region of Environment Canada for g iv ing me access to weather information. x v Chapter 1 Introduction The subject of this thesis is a study of the sea surface temperature (SST) off the northern end of Vancouver Island. This work is a contribution to a project on the Canadian west coast sockeye salmon behavior called: "Biophysical Controls of Salmon Migrat ion and Production". This project seeks to advance the capability to predict the migration route and run-timing of adult sockeye salmon returning to the Fraser River by looking at the influence of oceanographic variables and how they are influenced by the weather. 1.1 Motivation for this work The sockeye salmon spawn in the Fraser watershed, enter the marine environment in the spring of their second year, spend the third year into the open ocean and then return as mature fish in the summer of their 4th year (Hamilton, 1985). Despite the regularity of this cycle, uncertainties in homeward migration stil l remain. O n their journey back to the spawning ground, the sockeye salmon have the choice between two different routes when they reach the northern end of Vancouver Island: the passage through Johnstone Strait and the Strait of Georgia or along the West coast of Vancouver Island and through Juan de Fuca Strait. The percentage of sockeye salmon migrating via the northern route is referred to as the northern diversion or Johnstone Strait diversion and has a significant impact on the commercial fishery. Time series of the northern diversion from 1953 to 1994 show a variation between 2% and 80% (Fig. 1). Before 1978, values were relatively 1 uniform with a mean diversion rate near 17%. However, after 1977, diversion rates were much larger wi th inter-annual fluctuations. Over the past few decades, more research has been done to f ind the mechanisms responsible for the northern diversion's fluctuations. Water temperature, salinity and currents were found among the most important physical environmental parameters to determine the fish migration routes over the northeast Pacific due to their impact on swimming effort, availability of food and predation. In 1960, Tul ly et al. proposed a correlation between the northern diversion and warm water advection from the south, especially fol lowing an E l N ino event. In this case, more fish tended to use the northern route due to either temperature preferences or food distributions. Dur ing the fol lowing year, Favorite (1961) postulated a correlation wi th an area of low sea surface salinity extending seaward from the Queen Charlotte Sound, which led Wickett (1977) to make a correlation wi th the springtime Fraser River discharge for the period 1953-77. He concluded that the proportion of Fraser River water discharged into the ocean northwest of Vancouver Island increased the rate of Fraser sOckeye migrating through Johonstone Strait. Thereafter, Mysak (1986) used long-term sea surface temperature observations to suggest a correlation between the northern diversion and long-term environmental trends. In particular, significant changes in the atmospheric circulation were found dur ing E N S O events due to large scale re-distribution of the atmospheric pressure, temperature and humidi ty field. Furthermore, such changes were generally reflected in high sea level, high sea surface temperature and low salinity along the west coast of Brit ish Columbia, especially just after the mature phase of strong E N S O events, which then had an impact on fish behavior. Indeed, large catches of B.C. sockeye salmon occurred in several E N S O years (Mysak, 1986). Hami l ton (1985) 2 found that the northern diversion is better correlated wi th long period SST trends rather than wi th SST at the time of the return migration. This result indicates a connection wi th the overlying atmospheric circulation. Then, Emery and Hami l ton (1985) suggested a correlation between warming events over northeast Pacific and the northward transport in the oceanic surface layer induced by the surface w ind stress. Further analysis of the northern diversion indicated that the SST at Kains Island was found to be the best predictor for the period 1978-1983 (Groot and Quinn, 1987). Later, Xie & Hsieh(1989) developed a non-linear regression model for predicting the northern diversion rate using the SST at Kains Island and the Fraser river runoff. Their model works relatively wel l , however, further improvements to salmon migration route forecasts may be possible by including offshore SST data, the latter being more representative of the sockeye salmon's marine environment than coastal ones. 1.2 Objectives The principal objective of this thesis consists of an examination of a new set of data provided by A E S buoys to gain a better understanding of the SST distribution off northern Vancouver Island. The selected buoys, covering offshore waters over Queen Charlotte Sound (Fig. 1), have provided SST data since 1990. A lso, a comparison is made wi th coastal data provided by selected lighthouse stations and available A V H R R satellite imagery. 3 A second objective of this thesis is to explain the daily SST variations. LeBlond & Thomson (1990) indicated in their study on the North Pacific Ocean Environment that surface waters off the coast of British Columbia are mostly influenced by local processes such as local winds, solar radiation and upwell ing. Whi le the w ind forcing and induced upwel l ing are the most efficient mechanisms at changing the temperature characteristics of coastal water, the variations in the solar radiation may also have some impacts. Therefore, SST, w ind and cloud cover time series are studied together. Fol lowing Emery and Hamil ton (1985) results, the SST are also compared wi th synoptic weather maps of the northeastern Pacific. 1.3 Plan of the Thesis The thesis is organized as follows: Chapter 2 contains a description of the study area from previous studies. It includes bottom topography, temperature and salinity characteristics, water currents and meteorological forcing. Chapter 3 is a detailed description of data over the study area, including results from visual inspection of time series, data source and processing. Chapter 4 presents results from statistical analysis including correlation and regression analysis. Chapter 5 describes the background theory related to factors influencing SST including the conservation of heat energy and the oceanographic influences on SST variations such as upwel l ing, turbulent mix ing and horizontal advection. Chapter 6 is a comparison between windstress and SST along wi th synoptic maps and associated weather patterns. Chapter 7 presents results from a satellite remote sensing analysis including oceanographic features and transects of 4 sea surface temperature over the Queen Charlotte Sound /northern Vancouver Island area. Finally, a conclusion is presented in chapter 8. 5 100 90 80 70 60 50 40 30 20-i o k 0 Northern(Johnstone) Diversion 1 : 1 r -<D CO c g m > a 1955 1960 1965 1970 1975 1980 1985 1990 Year Queen Charlotte Sound Queen <&(^Charlotte Strait Johnstone 'trait Fraser River -Rosario Strait JuandeFucaS ^Salmon Strait Banks Figure 1. a) Proportion of adult sockeye salmon returning to the Fraser River via the northern route (through Johnstone Strait) (data from IPSFC). b) Schematic representation of Migratory routes of adult sockeye salmon around Vancouver Island (From Groot and Quinn, 1987). 6 Chapter 2 Description Of The Study Area. The study area covers the Queen Charlotte Sound and northern Vancouver island including the fol lowing three lighthouse stations: Mclnnes, Egg and Kains islands and the fol lowing four buoys stations: 46207 (East Dellwood), 46208 (West Moresby), 46204 (West Sea Otter) and 46185 (South Hecate) (Fig. 2). 2.1 Background 2.1.1 Bathymetry (Fig. 3) The Queen Charlotte Sound is a broad area between Vancouver Island and the Queen Charlotte Islands. It lies east of the Pacific ocean and the continental rise and west of the mainland of British Columbia. The bathymetry of the Queen Charlotte Sound is relatively complex due to extensive shallow areas and deep trenches reaching 350-400 m deep into the continental shelf. The main bathymetric features are the fol lowing: • Three shallow banks: Cook, Goose Island and Midd le banks. • Goose Island trough, a deep trench lying between Cook and Goose Island banks and extending towards the Northeast Pacific between Vancouver Island and the mainland. 7 • Mitchells trough extending between Goose Island and Midd le banks. • Moresby Trough running from Cape St. James to Banks Island. Numerous coastal indentations and seaways close to the main channels, especially the Goose Island Trough, allow oceanic water to f low into the coastal regions. Near northern Vancouver Island, the isobaths run generally parallel to the coast. The continental shelf, extending seaward to the 200m depth contour, has a wid th of approximately 20km near Cape Scott and narrows to approximately 15km at Kains Island. Beyond the shelf, the depth drops quickly to more than 1500m. 2.1.2 Tides This thesis concentrates on the effect of wind-dr iven currents and, to some extent, on the heat budget, but does not consider the tides explicitly. Nevertheless, the tides are the most energetic phenomena over B.C. waters and should be discussed briefly to appreciate their effect on SST. The waters in the vicinity of northern Vancouver Island and Queen Charlotte Sound are dominated by the semi-diurnal (twice daily) tide (M2). The semi-diurnal t idal wave propagates around the northeastern Pacific in a counter-clockwise direction at a speed of approximately 200 m/s . The tidal currents are rotary in the open water of Queen Charlotte Sound and become rectilinear as they enter Hecate Strait. Typical peak M2 speeds over that area are 20 to 40 cm /s (Crawford, 1994). Dur ing spring tides, maximum speeds are near 50 cm/s but decrease to 8 approximately half this value during neap tides. Near the shores, the tidal currents become aligned wi th the channels, and in many cases, the flow is accelerated as it enters the narrow mouth of these channels. The tidal range varies from approximately 2.4 m across the mouth of the Sound to around 3.0 m just north of Aristazabal Island, but may reach up to 5.0 m at the head of some channels. Figure 4 shows the co-range and co-phase values for the semi-diurnal tide (Thomson, 1981). The tidal current can affect the SST by its friction on the bottom. The latter generates mixing in the water column and can destroy the stratification. A s a result, the SST tends to remain more uniform and cooler in areas of strong tidal mix ing compared wi th surrounding waters. 2.1.3 Surface Salinity (Fig. 5) The water properties of the Queen Charlotte sound result from a combination of freshwater discharge from the coast and more saline offshore oceanic water. A t the end of the winter, before the freshwater river discharge has started, the isohalines run generally parallel to the coastline with typical values near 31.5 along the mainland coast to 32.5 at the line joining the southern tip of the Queen Charlotte Islands to the northern end of Vancouver Island. In May , as the freshwater discharges begin, the salinity distribution shows the formation of a trough of lower salinities in the southeastern area near Calvert Island. This trough extends seaward just north of Vancouver Island. As the summer season progresses, the fresh water input increases wi th the most important discharge over the southeastern section. This is reflected in lower salinities of coastal waters and an intensification of the southeastern trough. M in imum values can be 2-3 lower in August than at the 9 beginning of the summer (Dodimead, 1980). Further south, surface waters along the west coast of Vancouver Island are more saline and show an annual cycle. A typical monthly summer value for Kains Island is near 32.0 in August, and can be 0.1-0.3 higher dur ing an upwel l ing event. Dur ing the winter monthly sea surface salinity can be as low as 28.5 (Webster and Farmer, 1976). 2.1.4 Surface Water Temperature (Fig. 6) The monthly sea surface temperatures vary mainly wi th the amount of incoming solar radiation and the input of fresh water runoff. The monthly average values range from approximately 8°C in May to 14°C in August. Records of sea surface temperatures (Dodimead, 1980) show a weak temperature gradient over the Queen Charlotte Sound in early summer. Max imum sea surface temperatures are generally found near Goose Island and the Midd le banks. Jardine et al. (1993) looked at several years of A V H R R infrared imagery from the N O A A satellites and noticed an area of warm water developing late in summer over Goose Island bank. This can be explained by solar heating and weak tidal mixing preventing cold bottom water to rise to the surface. Just north of Vancouver Island, SST are lower due to fresh water discharge from rivers, especially the Bella Coola and Wannock rivers (Fig. 3) (Leblond et al, 1983). Then, a band of cool water has been observed off the northern tip of Vancouver Island which may be the result of horizontal advection of colder northern waters by the southward flow (Ikeda and Emery; 1984, Fang and Hsieh, 1993). However, from EOF analysis of A V H R R satellite imagery off Vancouver Island, Fang and Hsieh (1993) found their third mode indicating upwel l ing along the shelf break as another possible source. 10 Further north, a cold plume forms near Aristazabal Island, extends to the south, then west between Goose Island and Middle banks. Crawford et al. (1995) suggest that wind-dr iven upwel l ing may be responsible for this cold plume, whi le Jardine et al (1993) attribute some of the cooling to mixing. Just south of Cape St. James an area of cold and dense surface water is caused by intense mixing (Crawford et al, 1995). 2.1.5 Subsurface Temperature and Salinity Dur ing the summer season, when the northwest component of the w ind prevails, the relative offshore water mass transport increases near the surface over the Queen Charlotte Sound which triggers a compensating inshore movement of deeper waters and allows oceanic water to approach the coastal regions. Below approximately 50m depth, the salinity is everywhere greater than 32.8 and increases wi th depth to near 33.6 which is characteristic of the oceanic halocline. Figures 7 and 8 show the vertical profiles of temperature and salinity at two different locations (stations A and C) within the Queen Charlotte Sound for the months of July, August and September. The station A is located on the western side of the study area near 129.5W/51.2N and represents water properties just off the Queen Charlotte Sound. The station C, located near 129.2W/51.9N, represents water properties in the centre of the Sound. From these graphs, the main features are a thin mixed or near-mixed surface layer and strong thermocline and halocline. Under weak surface mixing, the thermocline extends from near surface to a depth of 75-100 m at the station A and to 100-125 m at station C. The halocline extends from 1 1 near surface to 125-150 m at both stations. Below the thermocline and halocline, temperature decreases and salinity increases slowly with depth (Dodimead, 1980). 2.1.6 Currents The general circulation over northeastern British Columbia consists of an eastward f lowing current known as the Subarctic Current or West W ind Drift. This current crosses the Pacific Ocean between 45 and 5 0 ° N and divides near Vancouver Island. One part turns northward to become the Alaska Current and the second part shifts southward to form the California Current (Thomson, 1981) (Fig. 9). Over the study area, the main dynamical forcing for currents are tides and w ind systems, the latter being much more efficient at changing the temperature and salinity characteristics of the water (Freeland et al, 1984) through turbulent mixing and upwel l ing, but also less predictable. While turbulent mixing is not a dominant factor dur ing the summer months due to weaker atmospheric circulation and w ind forcing, upwel l ing favorable winds are often observed along the west coast of B.C. from A p r i l through the end of September. The summer near-surface current along a line from the north end of Vancouver Island to Cape St. James flows southward as a result of dominant northwesterly w ind forcing (Fig. 10). Freeland et al (1984) used a simple one layer model to show that the core of the southward current is the result of an interaction between variable bottom topography and the wind forcing, so that upwel l ing and cold waters should appear first near the shelf break, especially under northwesterly winds. 12 Over the Queen Charlotte Sound the surface flow is more variable as a result of frictional effects and there is a tendency for the flow to be parallel to the local isobaths. A complex bottom topography lies below a shallow layer of water resulting in a clockwise gyre around Goose Island bank which is thought to be caused by tidal rectification (Freeland et al, 1984). There is also strong residual tidal currents from the Sound near Cape St. James (Foreman et al, 1992).. Freeland et al raised the importance of local winds at 2 to 30 day periods for surface currents. 2.1.7 Meteorological Forcing a) Large Scale Pressure Patterns The large-scale atmospheric circulation in the central North Pacific is determined by two major persistent centres of action: the Aleutian Low south and east of the Aleut ian Islands and the sub-tropical high pressure of the North Pacific wi th the westerly Jet Stream running between them. Superimposed on this climatological pressure pattern are synoptic weather systems (highs, lows and fronts) responsible for relatively rapid w ind fluctuations. The low pressure systems (lows and fronts) . develop and travel along the westerly Jet Stream at a speed corresponding to approximately 50% of its speed. A typical weather cycle is near 2 to 3 days over northeastern Pacific. On the other hand, the high pressure systems (anticyclones) tend to be slower and may influence the weather for several days. 1 3 1. Fa l l /Winter Pressure Pattern Dur ing the cold season, up to Ap r i l , the Aleutian low lies near 45° lat. N / 1 7 5 0 long. W and the North Pacific H igh , relatively weak, has its center near 30-40° lat. off the Cali fornia coast (Emery and Hamil ton, 1985) (Fig. 11). The mean position of the Jet Stream is near 5 0 ° N . This pattern produces prevail ing southwesterly geostrophic surface winds from the subtropical Pacific Ocean into the Gulf of Alaska. The superimposed synoptic lows, most intense over this period, tend to remain wel l offshore and track into the Gulf of Alaska along the Jet stream, while the associated frontal systems move across the B.C. coast (Fig. 12). The intensity of the lows and winds is determined by the contrast between cold and warm sectors (temperature gradient), the amount of tropical or sub-tropical moisture being injected into the system and the relative stability of the airmass. These systems usually give southeasterly gale (34 to 47 knots) to storm-force winds (48 to 63 knots), and occasionally hurricane-force winds (64 knots and more) to the B.C. coast. However, when the dynamics allow a rapid deepening of the lows near the coast (more than 24 mb deepening at the centre of the low within 24 hours), east to southeast winds may occasionally reach 70 knots with gusts up to 100 knots (Marine Weather Hazards Manual , 1990). 2. Spr ing/Summer Pressure Pattern A s the spring arrives, the baroclinic zone retreats northward due to increased solar radiation over the northern hemisphere. The Pacific Northwest begins to feel the effects of the Cali fornian high pressure system, which strengthens, moves further north and deviates the storm track into the northern Gulf of Alaska. The ridge, 14 extending over the Gul f of Alaska and off the B.C. coast, reaches its ful l development by the month of July. Meanwhile the Aleut ian low has weakened significantly and retreated to the north so that it is no longer evident, especially in July (Fig. 13). From May to September, the main position of the Jet Stream is around 5 3 - 5 5 ° N . Since the contrast between warm and cold air is much smaller, the strength of the Jet Stream has decreased significantly from winter, leading to weaker storms moving over the B.C. coast. The offshore ridge tends to dominate the northeastern Pacific, which gives prevail ing northwesterly winds along much of the B.C. coast up to gale-force strength. However, wind direction and speed fluctuate when an occasional front tries to penetrate through the ridge. 3. Deviat ion From The Ma in Pressure Patterns In other situations, especially during the winter, an Arctic airmass may spread over the southern interior of the province and push very cold air down the mountain passes. This situation usually brings cold, dry weather over much of B.C. and produces strong winds (gale or storm-force) through the coastal inlets that may persist for several days. 1 5 b) Summer Regional Scale Pressure Patterns Over Queen Charlotte Sound 1. Lee Trough , Dur ing the summer, the usual pressure pattern is dominated by the offshore ridge and a trough of low pressure in the southern interior of Brit ish Columbia (Fig. 14a). The latter is a thermal trough which develops after many days of sunshine. In more detail, when the land becomes relatively warm over southeastern B.C., the overlying airmass gets lighter which triggers an upward vertical motion, divergence of air aloft and lower surface pressures. A t high levels into the atmosphere, the divergent f low of air causes north-northeasterly winds which subside on the west side of the coastal ranges. Then, these subsiding winds converge wi th the outflow winds produced by the offshore high pressure system to form a trough of low pressure (called lee trough) just west of Vancouver Island (Fig. 14b). When the trough is relatively strong and the offshore ridge builds, a band of strong ' northwesterly winds develops over the water just behind the trough. Strong winds, rising at times to gale force, are often observed just off the northern end of Vancouver Island. However, further south just along the coastline of west Vancouver Island, the winds remain generally light from the east at 15 knots or less. This pressure pattern may persist for many days unti l a low pressure system moves eastward from the Pacific, forcing the offshore ridge to get closer to the coast and cutting the heating source by spreading clouds. Under such circumstances, the lee trough breaks down and the northwest winds (if still present) become much weaker. The combined effect of a thermal trough and the offshore ridge is probably the most 16 suitable situation for upwelling-favorable w ind with speed occasionally up to 65 k m / h r . 2. Summer Fronts When a frontal system approaches Queen Charlotte Sound, winds are generally from the south or southwest except southeast in the waters close to the coast because of the topography (Fig. 14c). Behind the front, as the ridge of high pressure rebuilds along the coast, the pressure may rise quite rapidly (0.5 to 1 mill ibars per hour) which produces strong northwesterly winds over the water. These Northwesterlies can be quite strong in the spring or early summer (May, June) as the fronts stil l have some of the strength of winter storms. 1 7 Figure 2: The northwest coast of British Columbia including the Queen Charlotte Sound. The study area lies within the dashed lines and the coastline. The buoys' positions are indicated by stars sign (*). (From Crawford et al, 1995). 134 W 130 W 126 W Figure 3. Bathymetry of the northwest coast of B.C. including the study area. Depth are in metres. (From C. Hannah, 1992). 19 Figure 4. Co-range and co-phase values for the semi-diurnal tide. Tidal range (broken line) in meters, tidal phase (solid line) in degrees. Difference of 29° corresponds to time difference of 1 hr. (From Thomson, 1981). 20 I V 132' . 131* 130* 129* 128' Figure 5. Salinity distribution (ppt) at 3 m depth a) for May 3rd-28th, 1954, b) for June 29th-July 22nd, 1954 and c) for August 17th-September 9th, 1954. The arrows indicate the direction of flows. (From Dodimead, 1980). 21 c) Figure 6. Temperature distribution in degree Celsius (°C.) at 3 m depth a) for May 3rd-June 20th, 1954, b) for June 29th-July 22nd, 1954 and c) for August 17th-September 9th, 1954. The arrows indicate the direction of flows. (From Dodimead, 1980). 2 2 V Figure 7. Temperature and salinity structures for waters in Queen Charlotte Sound at station A. (Circled numbers indicate years of survey). (From Dodimead, 1980). 23 Figure 8. Temperature and salinity structures for waters in Queen Charlotte Sound at station C. (Circled numbers indicate years of survey). (From Dodimead, 1980). 24 / a \ C ° * a \ BERING V ^ . ^ ^ ^ Z \ -M*S Y WESTEBAI v " * 7 = = = A i » « M n C ^ * I* / SUBARCTIC y — — -/ •' SUBARCTIC or«£ — Subarctic Current West tMnd Drtft ' SUBARCTIC BOUNDARY ' North P»cfflc Currant Figure 9. Prevailing surface currents in the North Pacific (From Thomson, 1981). Figure 10. July-August average near surface currents over the Queen Charlotte Sound and southern Hecate Strait. (From Crawford et al, 1995). 2 5 Figure 11. Mean sea level pressure pattern for January (pressure values are in millibars, mb). The arrows indicate winter storm tracks. (From Env. Canada, 1991). 26 Figure 12. Mean sea level pressure (mb) pattern associated with a deep winter low over the Gulf of Alaska with the related frontal system approaching the B.C. coast. (From Env. Canada, 1991). 2 7 Figure 13. Mean sea level pressure (mb) pattern for July. The arrows indicate summer storm tracks. (From Env. Canada, 1991). 28 a) b) Figure 14. Typical summer sea-level pressure (mb) pattern, a) Summer pressure trough over southern B.C. interior, b) lee trough (arrows and wind barbs indicate direction of surface winds. Wind barbs also indicate strength of winds). (From Env. Canada, 1991). 29 Figure 14. continued, c) Summer front. 30 Chapter 3 Description of Data 3.1 Data sources: Meteorological and Oceanographic Data. 3.1.1 Lighthouse stations. Dai ly sea surface temperature observations have been made at different locations along the B.C. coast; data were provided by the Institute of Ocean Sciences. This thesis examines 5 years of data (from 1990 to 1994) collected from three lighthouse stations which are: Mclnnes, Egg and Kains Islands (Table 1, Fig 2). Dai ly observations are made within one hour before the time of high tide, light schedules and other duties permitting, and taken at a depth of 0.9 metre. The temperature is measured wi th a mercury thermometer with an accuracy of 0.3 °C. or better for ind iv idua l measurements. Station Latitude N Longitude W Kains Is. 50°27' 128°02' Egg Is. 51°15' 127°50' Mclnnes Is. 52°16' 128°43' Table 1. Lighthouse geographic locations. There were only very few missing data (Table 2) that have been fi l led by linear interpolation. N o spike was found since the data set has already been checked at IOS. 3 1 Lighthouse station Year Missing days (day number/date) Number of missing data Kains Island 1990 none 0 1991 none 0 1992 116 (August 24th) 1 1993 none 0 1994 none 0 Mclnnes Isld. 1990 94 (August 02nd) 1 1992 55-57 (June 24th-26th) 3 1993 none 0 1994 none 0 Egg Island 1990 6,13 (May 06th,13th) 2 151 (September 28th^ 1 1991 05 (May 05th) 1 • " 118,121 (August 26th, 29th) 2 " 147 (September 24th) 1 1992 10 (May 10th) 1 " 43,52 (June 12th, 21st) 2 75 (July 14th) 1 102 (August 10th) 1 1993 67 (July 06th) 1 1994 17,27 (May 17th, 27th) 2 69 (July 08th) 1 Table 2. SST missing data. 3.1.2 Buoy data Four buoys from the Marine Weather Programs, Environment Canada -Pacific & Yukon Region, have been selected for the present analysis. They are located in the Queen Charlotte Sound, west of the Queen Charlotte Islands, off Cape Scott and over southern Hecate Strait (Table 3, Fig. 2). Station name Station ident. Latitude Longitude Water depth(m) East Dellwood C46207 50.8°N 129.9° 2125 West Moresby C56208 52.5° 132.7° 3000 West Sea Otter C46204 51.4° 128.8° 244 South Hecate St. C46185 52.4° 129.8° 220 Table 3. Buoy geographic location. 3 2 Hour ly buoy data used in this analysis are water temperatures, w ind speed and w ind direction. Water temperature is measured by a YSI 703 Linearized thermistor wi th an accuracy of 0.1 °C. Wind direction and wind speed are measured by a R M Young 5103 sensor wi th an accuracy of +/- 5 degrees and 0.6 m/s . Weather and ocean information was not reported at the buoy stations on a regular basis dur ing the study period, especially for the first two years. A lso, hourly water temperature below 5 °C and above 18°C have been considered as errors and removed from the data set. Gaps shorter than 10 percent of the time series have been fi l led by linear interpolation. Time series where gaps were between 5 and 10 percent must be interpreted wi th care. In some cases, the ful l time series are not available (see Table 4 for missing data). Missing data are relatively numerous, especially in 1990 and 1991, which does restrict the analysis. 3 3 Buoy station Year Missing days (day number, date) Nbr of missing days 46207 1990 29 (May 29th) 1 103-107 (August llth-15th) 5 62-72 (July 04th-llth) 11 1991 Series not available too many 1992 none 0 1993 none 0 1994 29 (May 29th) 1 71 (July 10th) 1 128,148 (September 05th, 25th) 2 46208 1990 Series not available too many 1991 1-61 (May Olst-Jule 30th) 61 1992 none 0 1993 1-24 (May 01st-24th) 24 1994 29 (May 29th) 1 46204 1990 29 (May 29th) 1 " 58-72 (June 27th-July 11th) 15 M 103-107 (August llth-15th) 5 1991 129-139 (September 06th-16th) 11 1992 none 0 1993 none 0 1994 • 29 (May 29th) 1 71 (July 10th) 1 82 (July 21st) 1 " 128,148 (September 05th, 25th) 2 46185 1990 Series not available too many 1991 61 (June 30th) 1 1992 none 0 1993 none 0 1994 29 (May 29th) 1 m 71 (July 10th) 1 m 128,139 (September 05th, 16th) 2 m 148 (September 25th) 1 Table 4. SST missing data. 3 4 Prepared hourly w ind data have been obtained from Josef Cherniawsky, IOS. Hour ly data obtained from the Pacific Weather Centre in Vancouver have been checked for suspicious records and gaps shorter than 2 days have been fil led by linear interpolation. 3.1.3 Other data Synoptic weather maps (12Z standard time) for northeast Pacif ic/western Canada and cloud cover data for three lighthouse stations (Cape Scott, Mclnnes Island and Cape St. James, See Fig. 2) have been provided by the Pacific Weather Centre, Environment Canada. C loud cover data available every 3 hours, except every hour for Cape St. James, have been converted into daily data for three specific events analyzed in chapter 6. 3.2 Data descriptions 3.2.1 Temperature 3.2.1.1 Variation In Time. Time series of daily sea surface temperature have been plotted for 5 consecutive summers (day l (May 1st) to day 153(September 30th) from 1990 to 1994) for the fol lowing lighthouse stations: Kains, Mclnnes and Egg Islands, and buoy stations: 46207,46208,46204 and 46185 (Fig. 15). 3 5 A summer thermal cycle in the upper ocean has been observed by Webster and Farmer (1976) at these lighthouse stations and is evident from SST time series wi th the seasonal maximum SST reached in late summer. Time series show a temperature increase generally un t i l approximately day 108 (mid-August) fol lowed by a slow temperature decrease to the end of the season. The fact that the SSTs begin to fall when the sea is still gaining heat has been explained by advective and /o r eddy processes (Tabata, 1958). The highest seasonal mean SST (14.0°C) was recorded in 1990 at the buoy 46204 while the lowest one (11.3°C) happened in 1991 at Egg Island (Table 5). From time series, daily fluctuations tend to have bigger amplitudes after day 75 (mid-summer). Pronounced variations occurred in 1992 after day 75 at the buoys 46185 and 46208, in 1993 after day 63 at the buoy 46185 and in 1994 after day 78 at the buoy 46185. The data have been checked, but no error has been detected. From seasonal mean SST, the warmest year appears to be 1994, while the coolest one corresponds to 1991. A previous study has shown that during an E l Nino event daily sea surface temperatures off the southern west coast of Vancouver Island may increase by as much as 3°C above normal (Freeland, 1990). However, the seasonal mean SSTs computed in this thesis for the years 1990 to 1994 show that the maximum values do not coincide wi th the last two E l N ino events (1991 and 1992-93) at any of the stations. 3 6 May 1st-Sept 30th. T.Min. (°C) T.Max (°C) T.Mean (°C) St.Dev. (°C) Kains Island. 1990 9.8 15.7 13.0 1.6 1991 9.5 14.8 12.1 1.3 1992 10.2 16.8 12.8 1.2 1993 10.0 15.7 12.7 1.4 1994 10.1 17.5 13.4 1.9 Mclnnes Island. 1990 8.3 16.3 12.5 1.8 1991 8.2 14.9 11.8 1.9 1992 8.8 15.5 11.8 1.4 1993 8.6 15.3 12.3 1.4 1994 9.3 15.7 12.8 1.6 Egg Island. 1990 8.8 15.5 11.6 1.4 1991 8.9 14.1 11.3 1.1 1992 9.7 15.2 11.7 1.1 1993 8.9. 14.7 11.8 1.3 1994 8.9 15.5 11.7 1.2 East Dellwood (46207). 1990 9.0 17.7 13.8 2.6 1991 — 1992 10.3 15.4 13.2 1.5 1993 9.0 15.9 12.9 1.7 1994 9.3 16.6 13.6 2.2 West Moresby (46208). 1990 — — — 1991 — — — — 1992 9.4 15.3 12.7 1.7 1993 — 16.2 — — 1994 9.0 16.3 13.0 2.3 South Hecate (46185). 1990 — — — 1991 8.3 15.9 12.5 2.0 1992 9.7 15.9 13.1 1.7 1993 8.8 16.1 13.2 1.8 1994 9.3 16.2 13.1 2.0 West Sea Otter (46204). 1990 9.3 17.2 14.0 2.2 1991 8.7 15.5 12.5 1.8 1992 9.5 15.0 12.6 1.5 1993 8.9 16.1 13.2 1.7 1994 9.5 16.5 13.3 1.8 Table 5. Statistics for SST (minimum, maximum, mean and standard deviation for the summer). 3 7 3.2.1.2 Variations In Space. Isotherms have been drawn from the mean SST distribution at every 0.5°C (Fig. 16) for each summer. The number of data points is certainly not sufficient to get detailed SST distributions. However the fol lowing features seem to emerge: A gradient of temperature perpendicular to the coast with relatively low values along the coast. The lowest SST are shown near Egg Island while the warmest are revealed by the buoys over Queen Charlotte Sound, especially at buoy 46207. Similar information can be seen from Table 5: Seasonal means and summer maximum SST are generally higher at the buoy stations than at the lighthouses. The summer maximum values were not observed simultaneously at al l stations, although the seasonal cycle is assumed to be the same over the scale of the study area for any particular year. The difference is relatively small between Kains and Mclnnes Islands (1 to 5 days), but can be as much as 15 days between Kains and Egg Islands. The buoy stations show as many variations, except for 1993 when summer maxima happen within 3 days for all buoys. The standard deviation of SST varies from 1.0°C to 2.1°C. Egg Island shows the smallest variations around the mean temperature, while SST fluctuations are more pronounced at the buoy 46185. 3.2.2 Synoptic And Wind Analysis Wind data at the buoy stations are available for the years 1990 to 1994. However due to numerous missing points, a relatively complete data set is only available for the last three years (1992, 1993 and 1994). The fol lowing analysis combines observations from time series and synoptic maps. 38 3.2.2.1 Variations In Time The months of May and June (day 1 to day 61) are usually transition months between the winter w ind regime and the summer one. Dur ing the first part of the summer winds are often light and variable. In the second part of the summer, when the high pressure system is wel l established over the northeastern Pacific, stronger northwesterly winds become dominant along the west coast of Brit ish Co lumbia . a) Summer 1990: The summer of 1990 is typical with variable winds from day 1 to approximately day 61 and mostly northwesterly winds after day 73 (Fig. 17a). The seasonal mean w ind speed, averaged over data from two buoys (46204 and 46207), is near 5.2 m / s from the west-southwest and the seasonal maximum, recorded at buoy 46207 on day 32, is 13.1m/s from the east-southeast (Table 6). From synoptic maps (Fig 18a), a north-south ridge of high pressure located approximately 450 km offshore on day 69 (July 8th) gave rise to northwesterly winds over part of the area. By day 73 (July 12th) the Northwesterlies were established over much of the B.C. coast. On day. 75 (July 14th) the pressure gradient strengthened over Queen Charlotte Sound as, a result of a thermal trough over southern B.C. interior and eastern Washington State and a low pressure system approaching from the west into the Gulf of Alaska (Fig. 18b). Northwesterly winds reached approximately 10-13 m / s (20-25 knots) over Queen Charlotte Sound on that day. Thereafter, the pressure pattern was dominated by a thermal trough over southern B.C. interior and a ridge of high pressure off the B.C. coast except for a 3 9 front that moved across the area on day 77 (July 16th). As a result, northwesterly winds prevailed over Queen Charlotte Sound unti l day 93 (August 1st). Between day 94 (August 2nd) and day 114 (August 22nd), a series of frontal systems moving across the area gave variable winds. For example, Figure 18c shows a frontal wave just off the northern end of Vancouver Island moving eastward. The pressure gradient on that day (day 94/August 2nd) was relatively weak wi th southeasterly winds generally less than 10 knots (5 m/s) in the vicinity or the northern end of Vancouver Island and along the eastern boundary of Queen Charlotte Sound. Behind the wave, winds were west to southwest near 10 knots (5m/s). The Nor th westerlies were re-established along much of the B.C coast on day 115 (August 23rd) due to an offshore high pressure system (Fig. 18d). These winds were maintained unti l the end of the summer (day 153) with variable speeds generally below 13.0 m/s . However, winds were variable at times, especially at the end of August and the beginning of September, as two weak fronts moved across the B.C. coast. b) Summer 1991: For 1991, time series (available only from the buoys 46204 and 46185) show relatively strong winds oscillating between northwesterly and southeasterly dur ing almost the entire season (Fig, 17b). The period of oscillation varied approximately between 4 to 12 days. The spatially averaged seasonal mean wind speed is around 4.9 m / s from the west-southwest and the maximum, recorded at the buoy 46185 on day 122, is 12.8 m / s from the south-southeast (Table 6). 4 0 The synoptic pattern for the summer of 1991 was slightly different from the seasonal mean: there were a few relatively strong frontal systems that moved across the B.C. coast dur ing the summer which kept the Californian ridge further south and gave relatively strong southeasterly winds along much of the B.C. coast. The most significant frontal systems moved through the area during the fol lowing times: between days 9-16 (May 9th-May 16th), 39-42 (June 8th-June 11th), 60-64 (June 29th-July 3rd) and 118-128 (August 26th-September 5th). Between these systems, the ridge off Cali fornia was able to bui ld over the northeastern Pacific and a thermal trough to develop over southern B.C. interior and Washington State. Both systems were responsible for northwesterly winds along the B.C. coast, especially from day 17 to day 27 (May 17th to May 27th) and from day 104 to day 116 (August 12th to August 24th). In particular, Figure 18e shows a thermal trough getting established on day 17 (May 17th) whi le the offshore ridge is relatively strong near 1 4 0 ° W . Thereafter, the thermal trough deepened into the interior of B.C. and the ridge moved towards the coast so that the pressure gradient and the Northwesterlies increased over Queen Charlotte Sound. However, by day 27 (May 27th) (Fig. 18f) the ridge was forced onto the coast due to an approaching frontal wave from the west, which ended the northwesterly event. A second northwesterly event began near day 104 (August 12th) as a thermal trough developed over southeastern B.C. (Fig. 18g). Meanwhile, the Cali fornian ridge extended across the Queen Charlotte Islands. As a result, Northwesterlies were relatively strong over Queen Charlotte Sound and persisted for several days. However, by day 116 (August 24th), the Californian ridge was located further away offshore, the thermal trough was fi l l ing (Fig. 18h) and the Northwesterlies began to weaken. Shortly after (day 118/August 26th), a strong frontal wave moved across the Queen Charlotte Islands, followed by a series of waves. 41 c) Summer 1992: Dur ing the summer of 1992, northwesterly winds were relatively frequent and also stronger (Fig. 17c). However, the seasonal maximum wind speed of 17.8 m / s is from the south-southeast and recorded at the buoy 46185 on day 151. The spatially averaged seasonal mean wind speed is near 5.5 m / s from the west (Table 6). Time series show three main northwesterly (upwelling-favorable) w ind events. The first one was observed early in the season between day 10 and day 21 (May 10th and May 21st) as a result of a low pressure system over Central Alberta wi th a trough extending along the Alaska panhandle and a broad high pressure system far offshore. There was also a thermal trough over southeastern B.C. after day 15 (May 15th). The associated northwesterly flow along the northwest coast generated w ind speeds up to 11 m / s near the northern end of Vancouver Island. Figure 18i shows the pressure pattern for day 10 (May.10th) wi th a trough extending over northern B.C. and a northwest-southeast ridge off the B.C. coast, both responsible for the northwesterly winds over much of the B.C. coast. On day 14 (May 14th), there was a thermal trough relatively wel l developed over southeastern B.C. whi le the offshore ridge had expanded over northeastern Pacific (Fig. 18j). Then, winds were relatively aligned wi th the coast (upwelling-favorable), especially in the northern'part of Queen Charlotte Sound and around the Queen Charlotte Islands. W ind speeds were up to 15 knots (7.8 m/s) especially north o f . 5 1 ° N . Thereafter, a series of frontal systems moved over northeastern Pacific and the B.C. coast during the last part of May and much of June (Day 23 (May 23rd) to day 55 (June 24th)). Another ridge of high pressure started to form offshore on day 62 (July 4 2 1st) (Fig. 18k), but winds remained light and variable over much of the study area for the next few days. O n day 78 (July 17th), the offshore ridge built stronger and a thermal trough developed over southeastern B.C. (Fig. 181). Further west, there was a low pressure moving into the Gulf of Alaska, increasing the pressure gradient along the B.C. coast. On day 78 (July 17th) northwesterly winds up to 11 m / s were recorded over much of the area. This northwesterly event ended on day 85 (July 24th) as a deep low pressure system moved into the Gulf of Alaska forcing the ridge onto the B.C. coast (Fig. 18m). Thereafter, a series of frontal systems moved over the B.C. coast giving variable winds over much of the coast. The last northwesterly event was initiated by a ridge that started to develop across the Queen Charlotte Islands dur ing the second week of August (day 104/August 12th), combined wi th a thermal trough over southern B.C. interior and Washington State (Fig. 18n). However, it is not unti l day 110 (August 18th) that northwesterly winds were relatively strong over much of the area, when the ridge moved just north of the Queen Charlotte Islands in the wake of a cold front (Fig. 18o). From day 115 to day 139 (August 23rd to September 16th) northwesterly winds were interrupted every 4 to 5 days by frontal waves moving across Queen Charlotte Sound. d) Summer 1993: W ind time series for the summer of 1993 are similar to those of 1990 (Fig 17d). Variable winds prevailed from day 1 to day 65 (May 1st to July 4th). From day 66 (July 5th) to almost the end of the summer relatively strong, sustained northwesterly winds were maintained over the area, especially from day 66 to day 81 (July 5th to July 20th) and from day 120 to day 145 (August 28th to September 22nd). The spatially averaged seasonal mean wind speed is 5.1 m / s from the west and the 4 3 maximum, recorded at the buoy 46185 on day 124, is 13.2 m / s from the north-northwest (Table 6).. Ridges and lows alternated every 4 to 5 days during the first part of the summer, explaining the variable winds mentioned above. For example, there was a weak low pressure area covering much of B.C. on day 35 (June 4th) giving light and variable winds along much of the coast (Fig. 18p). A few days later (day 38/June 7th), the low had moved along the B.C. and Alberta border, leaving a relatively strong ridge of high pressure near 1 3 5 ° W (Fig. 18q). As a result, the Northwesterlies up to 15 knots (7.8 m/s) were reported over Queen Charlotte Sound. However, on day 40 (June 9th), an intense frontal wave had replaced the ridge and the winds shifted to southeast (Fig. 18r). Dur ing the second part of the summer, the synoptic pattern was dominated by a quasi-stationary and strong high pressure system over northeastern Pacific (Fig. 18s) giving northwesterly winds up to 15 knots (7.7 m/s) along much of the B.C. coast. This ridge was also combined at times with a thermal trough over the southern interior of B.C. and a lee trough along the west coast of Vancouver Island. Figure 18t shows these features, wi th additionally a trough of low pressure extending along the north coast of B.C. and the Alaska panhandle. e) Summer 1994: W ind time series for the summer of 1994 look similar to the one of 1991 (Fig. 17e). A northerly w ind component alternated with a southerly one wi th a period varying generally between 4 to 15 days. Optimal conditions for upwel l ing were met between day 50 and day 86 (June 19th and July 25th). The spatially averaged w ind speed is 4 4 near 4.8 m / s from the west-southwest and the maximum w ind , recorded at the buoy 46204 on day 44 (June 13th), is 13.9 m/s from the east-southeast (Table 6). The synoptic pattern cannot be described as maps for 1994 are unavailable. Buoy Year Mean Wind Max Wind Day for max wind speed Misg. data on interval: Speed (m/s) Direction (degree) Speed (m/s) Direction (degree) 46204 1990 4.72 11.2 10.98 17.8 6 58-72 103-107 1991 4.66 30.0 12.61 154.9 100 129-139 1992 4.88 3.6 14.42 142.9 151 1993 4.34 -3.9 11.44 127.6 5 — 1994 4.32 33.3 13.85 158.0 44 — 46207 1990 5.64 9.3 13.12 147.8 32 64-72 103-107 1991 misg 1992 5.60 -11.0 14.41 107.7 151 — 1993 5.25 2.0 11.32 110.0 5 — 1994 4.95 10.4 12.06 130.6 31 — 46185 1990 misg 1991 5.07 33.3 12.78 125.8 122 — 1992 5.64 15.9 17.80 122.2 151 — 1993 5.14 -0.9 13.22 • -62.5 124 — 1994 4.96 38.0 12.39 142.6 28 46208 1990 miss 1991 misg 1992 6.06 13.3 13.75 -62.3 112 — 1993 5.67 -10.4 12.86 -57.8 97 1-24 1994 4.90 14.8 11.56 -65.6 83 — Table 6. Statistics for wind data. Wind direction (degree) indicates where the wind originates. It increases from 0° (westerly) to 90° (southerly) and to 180° (easterly). Negative wind direction decreased from 0° (westerly) to -90° (northerly) to -180° (easterly). See the following sketch. 4 5 3.2.2.2 variations in space Winds at the buoys tend to be highly correlated since they are located relatively far from topographic influences. Also, distances between buoys are small compared wi th the synoptic scale (scale of weather systems). V isua l inspection of w ind time series show relatively strong correlations among the four buoys for both, speed and direction. The correlation is particularly strong between buoys 46207 and 46204, although wind speeds tend to be somewhat stronger at the buoy 46207. Winds tend to be a little stronger at buoys 46208 and 46185, especially when the direction is from the northwest or southeast due to the orientation of the land and the fetch. 4 6 SST - Summer 1990 SST - Summer 1991 20 40 60 80 100 120 140 Day (Kains Island) 20 40 60 80 100 120 140 Day (Mclnnes Island) 40 60 80 100 120 140 Day (Egg Island) 40 60 80 100 120 140 Day (Buoy 46207) 20 40 60 80 100 120 140 Day (Buoy 46204) 16 O 14 •8 1 2 10 40 60 80 100 120 140 Day (Kains Island) 40 60 80 100 120 140 Day (Mclnnes Island) 40 60 80 100 120 140 Day (Egg Island) 40 60 80 100 120 140 Day (Buoy 46204) 40 60 80 100 120 Day (Buoy 46185) 140 20 40 60 80 100 120 140 Day (Buoy 46208) Figure 15. Sea surface temperatures in degree Celsius (°C) at lighthouse and buoy stations for the summers (day 1 (May 1st) to day 153 (September 30th)) of 1990 to 1994. See Table 4 for missing data. 47 SST - Summer 1992 SST - Summer 1993 20 40 60 80 100 120 140 Day (Kains Island) 20 40 60 80 100 120 140 Day (Mclnnes Island) 40 60 80 100 120 140 Day (Egg Island) 20 40 60 80 100 120 140 Day (Buoy 46207) 20 40 60 80 100 120 140 Day (Buoy 46204) 20 40 60 80 100 120 140 Day (Buoy 46185) 20 40 60 80 100 120 140 Day (Buoy 46208) 40 60 80 100 120 140 Day (Kains Island) 20 40 60 80 100 120 140 Day (Mclnnes Island) 40 60 80 100 120 140 Day (Egg Island) 20 40 60 80 100 120 140 Day (Buoy 46207) 20 40 60 80 100 120 140 Day (Buoy 46204) 20 40 60 80 100 120 140 Day (Buoy 46185) 20 40 60 80 100 120 140 Day (Buoy 46208) Figure 15 (continued) 48 SST - Summer 1994 40 60 80 100 120 140 Day (Kains Island) 20 40 60 80 100 120 140 Day (Mclnnes Island) 40 60 80 100 120 140 Day (Egg Island) 40 60 80 100 120 140 Day (Buoy 46207) 20 40 60 80 100 120 140 Day (Buoy 46204) 20 40 60 80 100 120 140 Day (Buoy 46185) 20 40 60 80 100 120 140 Day (Buoy 46208) Figure 15 (continued). 4 9 53 Z 5 2 ri> • § 5 1 Geographic Location CD <5 "^50 49 2oei \ ^ \: M • •207 204' E 34 -132 -130 -128 -126 Longitude (deg.W) Summer 1991 53 •52 £ 5 1 CD " § 5 0 49 m : w > 12.£ • m / ))l3 i2.A 34 -132 -130 -128 -126 Longitude (deg.W) Summer 1990 53 -52 © 51 CD "§50 CO 49 m : m~~l2.5 ( : i 3 i - V \ V^_^-^;13.0l f-4 34 -132 -130 -128 -126 Longitude (deg.W) Summer 1992 53 •52 CD "§50 CO 12. 7 r~~^\« \ * (12e N^13.2\ L 12\8^s 49 48 -134 -132 -130 -128 -126 Longitude (deg.W) Figure 16^  Distribution of seasonal mean SST (°C) for summers 1990 to 1994. Isotherms are drawn every 0.5°C. Missing data are indicated by "m". Relatively warm waters are indicated by "w". 50 Summer 1993 CO -•49 481 : : : ^ --134 -132 -130 -128 -126 Longitude (deg.W) Summer 1994 48 '• : : -134 -132 -130 -128 -126 Longitude (deg.W) Figure 16 (continued). 10r i : u T 1 1 r i 1 r n 1 1 1 1 r o £ -5[ -101 J L J l_ J l_ J I I-10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46207) 10r 1 1 -i 1 1 1 1 1 1 1 1 1 1 r 0 CD _ , > -5 -101 J L J L i ' I I I 1 1 L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46204) Figure 17. Wind time series in metre per second (m/s) at the buoy stations a) for the summer 1990 (Day 1 (May 1st) to day 153 (September 30th)). Data for the buoys 46185 and 46208 are missing. 5 2 Figure 17 (continued) b) for the summer 1991. Data for the buoys 46207 and 46208 are missing. 5 3 1 0 r-i 1 1 1 1 1 1 1 r ~i r 1 5 £• o 8 S -5h -10 10 • i i • i * i i i J L J l_ J 1_ 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46207) 1 r 1 1 1 r-—7"r ML A \\A i 5 ] £ -5 -10 10 jo 5 £ 0 8 £ -5 "i 1 1 1 1 1 r J I L J L ' • I I I L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46204) T 1 1 1 1 1 1 1 1 1 1 r ~ i — n r -10 10 « 5 £ o 8 £ -5 -10 J L V J I L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) i 1 r i 1 r n 1 1 1 1 1 r J I I I I L 4 J I L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46208) Figure 17 (continued) c) for the summer 1992, 5 4 10 r I 51 ¥ of 8 i 1 1 r T r -i r i r -101 • * i i i i i 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46207) 10r - i 1 1 1 1 1 r 1 1 1 1 1 r .» 5 i 0 £ -5\ -101 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46204) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) 10r | 5\ £ Of o * -5\ T 1 1 1 1 1 1 1 1 1 1 i i r XT' -10l I I I I i I I Hi 1 L _LJ L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46208) Figure 17 (continued) d) for the summer 1993, T 1 1 1 r r. i r r i r J L J I I I I I L 1 0 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46207) T 1 1 1 1 1 1 1 1 r - i i r 10 « 5 i - o 8 CD _ > -5 _L J I L I I I I I L " 1 ° 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46204) 10 ^ 5 i 0 CD -> -5 -10 V 1 1 i 1 1 1 1 1 m 1 i i r > i i J I : 1 1 L 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) CD _ > -5 -10 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46208) Figure 17 (continued) e) for the summer 1994. 56 b Figure 18. Synoptic weather maps for northeastern Pacific/western Canada with sea-level pressures in millibar, a) July 8th, 1990, b) July 14th, 1990, 5 7 Figure 18 (continued) c) August 2nd, 1990, d) August 23rd, 1990, 5 8 Figure 18 (continued) e) May 17th, 1991, f) May 27th, 1991, 5 9 Figure 18 (continued) g) August 12th, 1991, h) August 24th, 1991, 60 Figure 18 (continued) i) May 10th, 1992, j) May 14th, 1992, 61 62 Figure 18 (continued) m) July 24th, 1992, n) August 12th, 1992, 63 Figure 18 (continued) o) August 18th, 1992, 64 Figure 18 (continued) p) June 4th, 1993, q) June 7th, 1993, 65 Figure 18 (continued) r) June 9th, 1993, 66 Figure 18 (continued) s) July 5th, 1993, t) July 9th, 1993. 67 Chapter 4 Statistical Analysis 4.1 Regression Analysis A regression is applied to the time series of temperature to isolate and eliminate the low frequency seasonal signal. The residuals about a least square fit w i l l be called SST anomalies. Let the temperature at the sea surface be represented as follows: r = A + flsin(G») + Ccos( f l» ) + £ (4-1) where co = 2;r / 365 except co = 2KI'366 for the bissextile year (1992), t=time, A , B and C are the model's parameters; T represents sea surface temperatures (SST) and e represents SST anomalies about the fit. In order to f ind the values for A , B and C, I have used the least square fit method to minimize residuals. Therefore, we need to find the values of A , B and C so that the fol lowing expression: 153 £ e 2 = ^(T. -A- flsinCow,.) - Ccos(f i»,.)) 2 (4.2) 1=1 is m in imal . 6 8 4.1.1 Results From Regression Analysis The coefficients A , B and C, the amplitude, phase and root mean square of residuals (RMS) resulting from the regression analysis are presented in the fol lowing tables (Tables 7 to 13) for the three lighthouses and the four buoys over the summer season (from May 1st to September 30th). Figure 19a and 19b show examples of the annual cycle superimposed on the SST for Kains Island and the buoy 46207 between 1990 and 1994. The main difference between the two sites is the amplitudes of the SST variations, the latter being generally larger at Kains Island. The residuals (SST anomalies) are analyzed in more detail in the fol lowing section. From the mean annual SST (coefficient A) , 1994 was generally the warmest year wi th the highest value found at Kains Island. Isolines of constant amplitude (Fig 20) indicate larger variations of the annual cycle at the buoys 46207 and 46204, which also correlates wi th spatial distributions of mean SST presented in the previous chapter. 6 9 Regression Analysis: 1990 1991 1992 1993 1994 A 10.56 10.07 10.69 9.98 11.93 B 3.67 3.05 2.99 3.79 2.74 C -1.12 -0.74 -0.03 0.32 -1.71 Amp. 3.84 3.14 2.99 3.80 3.23 Phase -0.30 -0.24 -0.01 0.08 -0.56 RMS 0.66 0.61 0.92 1.00 1.00 Table 7. Kains Island. A 9.96 9.30 9.25 9.21 11.25 B 3.88 3.94 3.66 4.32 2.55 C -1.13 -1.54 -0.64 0.24 -1.54 Amp. 4.05 4.23 3.72 4.33 2.98 Phase -0.28 -0.37 -0.17 0.06 -0.54 RMS 0.89 0.57 0.67 0.91 0.78 Table 8. Mclnnes Island. A 8.58 9.22 8.92 8.72 9.13 B 4.32 2.99 3.74 4.13 3.56 C -0.11 -0.28 0.80 0.62 0.003 Amp. 4.32 3.00 3.82 4.18 3.56 Phase -0.03 -0.09 0.21 0.15 0.00 RMS 0.77 0.69 0.67 0.84 0.72 Table 9. Egg Island. 7 0 ;ression Analysis (continued): 1990 1991 1992 1993 1994 A 10.42 10.02 8.98 11.38 B 5.33 4.64 5.53 3.66 C -2.29 -0.55 -0.07 -2.29 Amp. 5.80 4.67 5.54 4.32 Phase -0.41 -0.12 -0.01 -0.56 RMS 0.58 0.44 0.72 0.53 Table 10. Buoy 46207. A 10.91 10.84 B . . . . 2.86 3.67 C -1.53 -2.44 Amp. 3.25 — 4.41 Phase -0.49 -0.59 RMS 0.80 0.49 Table 11. Buoy 46208. A 10.42 8.93 8.28 9.19 -— B 5.40 5.18 5.92 5.60 C -1.68 -0.83 0.67 -0.04 Amp. 5.66 5.25 5.96 5.60 Phase -0.30 -0.16 0.11 -0.01 RMS . 0.46 0.58 0.50 0.69 Table 12. Buoy 46204. A 9.94 9.95 10.12 10.57 B 4.00 4.66 4.57 3.98 C -1.84 -0.90 -1.08 -1.68 Amp. 4.40 4.75. 4.69 4.32 Phase -0.43 -0.19 -0.23 • -0.40 RMS 0.48 0.62 0.73 0.63 Table 13. Buoy 46185. 7 1 4.1.2 SST Anomalies Time Series Analysis (Fig 21) After removing the low frequency signal, visual inspection of time series show the most evident cycle common to all stations at 2 to 3 days (Fig. 21). This is suggestive of a relation wi th synoptic weather systems as discussed in chapter 2. Then, variations in SST anomalies are not wel l defined. They seem rather uncorrelated from one station to the other and from one year to the next. However, a weak cycle seems to emerge at 6 to 7 days and up to 10 days which also suggests a connection wi th the atmospheric circulation. Dur ing the summer, slow moving weather systems giv ing quasi-stationary weather conditions for several days are common along the west coast of B.C which may be responsible for this cycle. A long cycle of 40 to 60 days and up to half the summer period can also be seen in some time series, especially for the summer of 1993 and 1994 (Fig. 21d to 21e). In more detail, the time series show that high frequency fluctuations tend to have smaller amplitudes at the buoys than at the lighthouses, especially for the years 1990 and 1992. However, SST time series anomalies show episodic large amplitudes at the buoys 46185 and 46208 (see Fig. 21c,d and e). SST anomalies vary approximately between -2.0°C and 2.3°C at the lighthouses and between -1.7°C and 1.5°C at the buoys. Time series suggest similarities between the buoys in high and low frequency variations, especially for the buoys 46204 and 46207. By a careful inspection, one can see similarities between coastal and offshore stations, especially between Kains Island and the buoy 46204. For example, the 1993 data (Fig. 21d) show a correlation at 75 days relatively in phase for all stations with two crests near day 35 and day 110 and two troughs near day 70 and day 145. Low frequency oscillations also emerge from the 1994 time series, but do not appear correlated in space. For the years 1991 to 1992, such slow oscillations do not have much energy. However, correlations can be 7 2 found at higher frequencies. For example, the data for 1992 (Fig. 21c) indicate an oscillation at approximately 10 days which show a relatively strong correlation in space over much of the summer, especially between Kains Island and the buoy 46204. 4.2 Spectral Analysis The spectral analysis technique using the fast Fourier transform (FFT) has been performed on detrended time series (SST anomalies time series) by using a Hanning window oyer blocks of length of 64 days (M). From Priestley (1989), the bandwidth for a Hanning window is given by K 5 = 2.45 MAt With a sampling interval At equal to one day, the bandwidth is 0.12 cycle per day (cpd). The number of degrees of freedom is given by v = 2.45N IM, where N (153) is the number of observations. Therefore, there are approximately 6 degrees of freedom. 4.2.1 Results From Spectral Analysis The figures 22a, 22b and 22c present examples of power spectral density (PSD) for Kains Island and the buoys 46207 and 46204 over the study period (1990 to 1994). The 95% confidence intervals indicated by the dashed lines are estimated by calculating the variance of the unaveraged spectral estimates under the assumption of a normal 7 3 distribution of nonoverlapping sections. Provided this assumption is correct, there is a 95% probability that the confidence interval covers the true PSD (For more details, see Signal Processing Toolbox, User's Guide for Matlab). The main feature emerging from PSD is that there is generally more energy towards the low frequencies (red signal) compared with high frequencies. Most PSDs show a relative maximum near 14 days (0.07 cpd). This cycle may be due to the fortnightly t idal variation which affects the method of sampling. However, the weakness of these maxima indicates that other phenomena are also involved wi th in a range of frequencies close to 0.07 cpd. A t the other end of the spectrum, a cycle just above 2 days (near 0.44) is common over most stations and in most summers, which agrees wi th visual inspection (section 4.1.3). From a comparison between Kains Island and the buoys 46204 and 46207, there is also a 4 day cycle (0.25 cpd) showing up for almost every year (Fig. 22a, b and c). However, this energy peak is above the 5% level of significance (0.88) only for the year 1990 and 1994. 4.3 Correlation Analysis 4.3.1 Correlation Coefficient The correlation coefficient, a measure of intensity of association between two variables, is calculated as follows: (4.3) 7 4 where Tu and T2i are the temperatures for series 1 and 2 observed at time " i " , and Tx and T2 are the seasonal mean temperature for series 1 and 2. This coefficient does not have units and varies in the range -1 < r < 1. 4.3.2 Results From Correlation Analysis - SST anomalies The auto-correlation analysis for Kains Island reveals relatively short de-correlation scales from 1990 to 1992 (4 to 8 days/19 to 38 degrees of freedom(DF)) and longer ones for the years of 1993 and 1994 (near 18 days/8 DF). Mclnnes Island seems to have a longer memory wi th a de-correlation scale between 12 and 18 days (13 and 9 DF), except for the year 1992 where it is only 5 days (30 DF). Finally, the mean de-correlation scale is 7 days (21 DF) at Egg Island and varies generally between 6 and 18 days (8 and 25 DF) at the buoy stations. The correlation coefficients among buoys and lighthouses vary from nearly 0 to 0.76 wi th highest values generally reached in 1993 (tables 14 to 18). The highest correlation between time series seems to be revealed by the buoys 46204 and 46207 (0.76) of that year, while nearly no correlation is shown by Mclnnes Island and the buoy 46204 (0.01) in 1990. However, from the 95% confidence intervals, only eight coefficients are significantly different from zero (marked wi th star sign in Table 14 to 18). Most of them are revealed by the buoy data in 1992 and 1993. Among the lighthouses, the highest significant correlation coefficient (0.47) is shown by Kains and Mclnnes Islands in 1992. A comparison between Kains Island and the buoy 7 5 coefficients shows the highest significant correlation wi th the buoy 46204 in 1993 (0.61). Correlation coefficients(r) for SST time series: 1990 1991 1992 1993 1994 Kains/Mclnnes 0.40 0.48 0.47 * 0.63 0.58 Kains/Egg 0:10 0.35 0.15 0.39 0.08 Mclnnes/Egg -0.30 0.18 0.32 0.53 0.28 Table 14. Lighthouses 46208/46207 — — 0.23 — 0.66 46208/46185 — — 0.52 * — 0.46 46208/46204 — — 0.22 • — 0.35 46207/46185 — — 0.59 * 0.53 0.28 46207/46204 0.39 — 0.55 * 0.76 * 0.35 46185/46204 — 0.59 0.47 0.67 * 0.56 Table 15. Buoys. 46208 — — 0.13 — 0.15 46207 0.38 — 0.32 0.56 -0.07 46204 0.20 0.46 0.17 0.61 * 0.48 46185 — 0.26 0.19 0.47 * 0.36 Table 16. Kains Island vs buoys. 46208 — — 0.24 — 0.12 46207 0.06 — 0.53 0.58 0.08 46204 0.01 0.17 0.35 0.66 0.51 46185 0.07 0.38 0.55 0.38 Table 17. Mclnnes Island vs buoys. 46208 '-.— — 0.09 — 0.27 46207 0.20 — 0.16 0.34 0.29 46204 0.22 0.35 0.34 0.43 0.41 46185 — 0.30 0.00 0.34 0.13 Table 18. Egg Island vs buoys. 7 6 4.3.3 Results From Correlation Analysis - Alongshore Wind Component The alongshore w ind is obtained by rotating the u and v-components by 45° . Then, the correlation coefficients have been computed for the alongshore w ind component. Numbers are relatively high (Table 19), as expected from visual inspection of w ind time series, and they are all significantly different from zero. The lowest values (near 0.60) are found between the buoys 46208 and 46204, whi le the highest ones (near 0.90) involve the buoy 46207, 46204 and 46185. The de-correlation scales for w ind time series (between 3 to 8 days) is generally shorter than the ones for SST time series. This is not surprising due to the great inertia of the ocean compared to that of the atmosphere. 1990 1991 1992 1993 1994 46207/46208 0.81 0.61 0.73 46207/46204 0.89 0.91 0.86 0.90 46207/46185 * 0.92 0.85 0.86 46208/46204 0.76 0.60 0.62 46208/46185 0.80 0.75 0.78 46204/46185 0.82 0.88 0.88 0.87 Table 19. Correlation coefficients for alongshore wind time series at the buoys. 4.4 General Observations From most results in this chapter, it seems evident that statistical analysis does not give enough information about the temperature variations of the surface water. The main reason is the large number of variations caused by complex oceanic and /o r atmospheric processes at different scales, both in time and space. However, 7 7 by visual inspection of the time series, there is evidence of some correlations for certain periods of time. In order to understand the variations of the surface water temperature, three events are analyzed in chapter 6. Details are given in terms of SST compared wi th w ind and cloud data, along with synoptic pressure patterns. However, before we get into the analysis, a review of the relevant theory is presented in the next chapter. 7 8 _] I I 1 I I I I I I 1 I I I L_ 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (1991) T 1 1 1 1 1 1 1 1 I 1 1 1 1 T _l I I I I I I I I I I I 1 1 L J 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (1992) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 T i i I I I I I I I I I I I I L_ 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (1993) T 1 1 1 1 1 1 1 1 1 1 1—: 1 r 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (1994) Figure 19 (continued), b) for the buoy 46207. 80 Geographic Location Longitude (deg.W) Summer 1991 53 52 d) CD T 3 51 CD •o 50 CO _ l 49 48 - f c m; ; 4 4 0 4.2JV 34 -132 -130 -128 -126 Longitude (deg.W) Summer 1990 53 •52 O) a) " § 5 0 CO 49 4; ?34 -132 -130 -128 -126 Longitude (deg.W) Summer 1992 53 z 5 2 CD " § 5 0 49 4 I34 -132 -130 -128 -126 Longitude (deg.W) Figure 20. Distribution of the amplitude of the annual cycle of SST (°C) for summers 1990 to 1994 (from regression analysis). Isotherms are drawn every 0.5°C. Missing data are indicated by "m". Relatively warm waters are indicated by "w". 81 Summer 1993 Summer 1994 48' ; ; ; : 481 : ' • — -134 -132 -130 -128 -126 -134 -132 -130 -128 -126 Longitude (deg.W) Longitude (deg.W) Figure 20. (continued). 8 2 Figure 21. Sea surface temperature anomalies(°C) at lighthouse and buoy stations for a) summer 1990 (Day 1 (May 1st) to day 153 (September 30th)). The data for the buoys 46185 and 46208 are missing. 8 3 Figure 21. (continued) b) summer 1991. The data for the buoy 46207 are missing. 8 4 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) Days (Buoy 46208) Figure 21. (continued) b) summer 1991. 8 5 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Kains Island) Figure 21. (continued) c) summer 1992. 86 10 20 30 40 50 60 • 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) Days (Buoy 46208) Figure 21. (continued) c) summer 1992. 8 7 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Mclnnes Island) -I 1 1 1 I I I i I i i I I I l_ 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Egg Island) 1 i i i 1 1 1 1 1 1 1 1 1 1 r -J 1 1 1 1 1 1 1 l I I l i i i 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46207) Figure 21. (continued) d) summer 1993. The data for the buoy 46208 are missing. 88 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Days (Buoy 46185) Figure 21. (continued) d) summer 1993. 8 9 Figure 21. (continued) e) summer 1994. 90 Figure 21. (continued) e) summer 1994. 91 Figure 22. Power spectrum density for SST anomalies at a) Kains Island. The area within the dashed lines represents the 95% confidence intervals. The bandwidth is 0.12 cpd and the degrees of freedom is 6. 92 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Frequency (cpd) (1990) 10 8 co 6r o °- 4h 2h Missing Data 0.05 0.1 0.15 0.2 0.25 0.3 Frequency (cpd) (1991) 0.35 0.4 0.45 0.5 0.1 0.15 0.2 0.25 0.3 Frequency (cpd) (1992) 0.35 0.4 0.45 0.5 10-0.05 0.1 0.15 0.2 0.25 0.3 Frequency (cpd) (1993) 0.35 0.15 0.2 0.25 0.3 Frequency (cpd) (1994) 0.35 0.4 0.45 0.5 Figure 22. (continued), b) Buoy 46207, 93 Figure 22. (continued), c) Buoy 46204. 94 Chapter 5 Background Theory The first part of this chapter (section 5.1) presents an overview of the main factors affecting the-SST along the west coast area of British Columbia based on past studies. In particular, it outlines the mechanisms affecting the local stratification which relates to the variabil ity of SST. The next section (5.2) is a review of the heat budget of the surface waters, covering the most important terms responsible for SST variations. Then, section 5.3 presents a summary of Ekman theory, relating the w ind forcing to the ocean's response in the upper layer. In particular, it includes a detailed description of the upwel l ing phenomena which appear as a dominant factor in modulat ing SST over much of the study area as shown in chapter 6. 5.1 Vertical Stratification The mechanisms responsible for the stratification of the shelf waters (within the 200m depth contour) fall under the following two categories: The first category involves mechanisms that create stability, i.e., the buoyancy forces resulting from freshwater input and surface heating. The second category includes processes that destroy it, i.e., the negative buoyancy forces resulting from evaporation and cooling and the mixing forces such as wind and bottom shear generated turbulence. Most of these mechanisms can lead to SST fluctuations wi th a time scale from a few to several days (2 to 10 days). River runoff is l ikely to be the primary source of buoyancy that creates stratification especially in the inner and middle shelf near rivers (Atkinson and Jackson, 1986). In 9 5 comparison wi th solar heating that may be equally important over the entire shelf area, Garrett et al. (1978) have shown that the buoyancy flux at rate R is equivalent to that due to a heat flux Q if R = aQcp-\Ap) where R is the fresh water input from rainfall, a is the thermal expansion coefficient of water, c p the specific heat and Ap the density difference between fresh and salt water. This relationship, in agreement wi th Atkinson and Jackson's results (1986), shows that the stratification may result from both fresh water input and heat fluxes in areas where rivers discharge is significant, assuming that R represents freshwater input from river discharge instead of rainfall. From historical salinity data (Dodimead, 1980), the presence of freshwater is revealed along the mainland coast, especially over the southern part of Queen Charlotte Sound during the summer. Figure 4 shows a tongue of low salinity along a northeast-southwest line just south of Calvert Island that was particularly pronounced from June 29th to July 22nd, 1954. Therefore, the freshwater input w i l l increase the stability of the water column in the vicinity of Cook bank. Wi th in a time scale of a few days, the surface wind was found to be a dominant factor in cooling surface waters, especially during intense Pacific storms (Large and Crawford, 1994). Their experiment, the "Ocean Storms Experiment", also revealed that the resulting w ind stress and entrainment process may cause significant cooling of the SST wi th in a single day by mixing the upper layer and entrainment across the thermocline, depending on the w ind speed, the local stratification and the Coriol is parameter. Dur ing the summer, w ind speeds are generally weak along the west 9 6 coast of B.C. due to a ridge of high pressure prevail ing over northeastern Pacific, especially in July and August. Therefore, the entrainment process and the cooling of the surface water caused by the vertical mixing are expected to be weaker during that season. However, northwesterly winds generated ahead of the ridge may cause significant cooling of the coastal surface water during the summer by induced upwel l ing (section 5.3.1). This can be seen from seasonal mean SST (Table 5) showing lower values toward the shore. Turbulent mixing caused by bottom friction and tidal currents is another mechanism that weakens the stratification by mixing the water column vertically. Simpson and Hunter (1974) proposed a parameter proportional to HI £ / 3 for the separation of mixed and stratified regions, where H is the depth of the water and U , the mean speed of tidal current. They assumed the heat flux, density, specific heat, thermal expansion and bottom friction coefficients uniform over an area. Based on their mix ing parameter S = \ogl0(H/ Q £ / 3 ) ) where C d is the bottom drag coefficient, Jardine et al (1993) produced maps of stratification distribution over the Queen Charlotte Sound and Hecate Strait for spring and neap tides (Fig. 23). Maps show stratified conditions over much of the study area except in the vicinity of Goose Island bank and at the northern end of Vancouver Island. 9 7 5.2 Heat Budget (Pickard and Emery, 1982) The amplitude and character of the SST variations depend on energy exchange processes occurring at the air-sea interface and advective processes occurring below the sea surface. Whi le the energy exchange processes include the solar radiation, effective back radiation, evaporation and conduction of sensible heat, the advective processes are concerned with transport of water resulting from wind- , t idal- and /o r density forcing. The net rate of heat flow at the air-sea interface can be stated by the fo l lowing equation Qs + Qb + Qh + Qe + QV = QT where Qx represent the different components of the heat budget as follows: Qs, the rate of inflow of solar energy through the sea surface, Qb, the net rate of heat loss by the sea as long-wave radiation to the atmosphere and space, Qh, the rate of heat loss/gain through the sea surface by conduction, Qe, the rate of heat loss/gain by evaporation/condensation, Qv, the rate of heat loss/gain by the ocean due to advection and QT is the resultant rate of gain/ loss of heat of the ocean. Dur ing the summer season, the terms, in order of importance, are Qs which is always positive, Qb and Qe which are always negative and Qh which is slightly positive. Advect ion Qv may be quite variable over space and time. When QT is positive (negative), there is gain (loss) of heat into (from) the ocean and an increase (decrease) of SST. The different terms of the above equation are discussed indiv idual ly below. 9 8 5.2.1 Solar radiation(<2s) The main source of energy for a column of water near the surface is the short-wave radiation emitted by the sun. When entering the atmosphere, parts of the incoming short-wave radiation are lost by scattering and absorption due to interaction wi th the atmosphere's particles and water vapor. However, since the atmosphere re-emits some energy downward (called sky radiation), the loss is reduced. From the energy reaching the sea, a small fraction is reflected at the surface and the rest is used to increase the SST. This energy flux is always positive, however it varies over space and time due to variations in the composition of the atmosphere and the degree of cloudiness. As a reference. Figure 24 (from Pickard and Emery, 1982) shows the daily inf low of solar radiation at the Earth's surface as a function of latitude and time of year, where no cloud and an average atmospheric transmission of 70% have been assumed. Near the latitude of the study area ( 5 1 ° N ) , the daily rate of short-wave radiation (Qs) varies from 200 W / m 2 in May to less than 100 W / m 2 in September. Max imum values estimated at 250 W / m 2 are found toward the summer solstice. A study by Tabata in the vicinity of Triple Island, B.C. (1958) indicates a summer mean value for Qs near 230 W / m 2 . The cloud cover is a major factor for the amount of solar radiation reaching the sea. Tabata (1958) showed that the incident solar radiation at the sea surface under mean cloud cover was approximately half of the incident solar radiation in the absence of clouds. Thus, year-to-year fluctuations of monthly means were mostly associated wi th variations in cloud cover. The presence of clouds in the atmosphere reduces parts of the incoming short-wave radiation by scattering and absorption due to interaction wi th cloud particles. To 9 9 take this effect into account, the mean energy which would arrive in the absence of clouds may be mult ipl ied by the fol lowing conversion factor: (1-0.09C) where C is the proportion of sky covered by cloud in eighths (oktas) (Pickard and Emery, 1982). It is noteworthy that this method neglects the sunlight scattered by the atmosphere and clouds. However, during the summer a large fraction of Qs is direct sunlight. A second factor neglected is the reflection at the sea surface which depends upon the elevation of the sun and the sea state. While the reflection on a flat sea can be estimated from a coefficient table, the wave effect is more difficult to estimate. 5.2.2 Long-Wave Radiation (Qb) The long-wave radiation term Qb, also called back radiation, reflects the net amount of energy lost by the sea as long-wave radiation. The net transfer of long-wave radiation is a function of the sea and air temperatures, the vapour content of the air and the cloud amount and height. The magnitude of Qb is proportional to the fourth power of the absolute temperature of the emitter; it is always negative, meaning that it contributes to decreasing the SST. This term can be measured with a radiometer as described in Pickard and Emery (1982). However, when direct measurements are not available, the heat loss by the sea may be estimated from a method taking into account the absolute SST and the overlying water-vapor content. While the SST gives an estimate of the rate of outward f low of energy, the water-vapor content provides a value for the inward f low from the atmosphere since the atmospheric water vapor is the main source of its long-wave radiation. Values for Qb may be read from Figure 25 when the SST and the relative humidi ty are known. For example, a relative humidi ty near 90% 1 0 0 with the SST at 12°C gives an approximate value for Qb of -92 w / m 2 , assuming no cloud. In the presence of clouds, the contribution from the atmosphere is increased so that the net loss of heat is reduced. This effect may be taken into account by using the factor (1-0.1C) where C is the amount of sky covered by clouds in oktas (Pickard and Emery, 1982). The back radiation term is assumed to be relatively constant in time and space because it depends on the absolute temperature, rather than the Celsius temperature, and because the relative humidity does not change much over the sea^ The results from Tabata (1958) show a net loss of heat from the effective back radiation near -82 g -ca l / cm 2 / day (-40 W / m 2 ) relatively constant from May to September. Therefore, under clear skies the effect of Qb on SST variations may be neglected. However, when clouds are present, the loss of energy as long-wave radiation can be sharply reduced. This may result in a noticeable fluctuation in heat flux from one day to the next. Therefore, variations in Qb could be important for those in the SST. 5.2.3 Conduction Heat may be gained or lost from the sea surface by conduction due to the temperature gradient in the air adjacent to the sea. When the air temperature decreases upward, heat is conducted away from the sea surface. The rate of loss of heat Qh is proportional to the heat conductivity coefficient K, the specific heat of air at constant pressure C p and the vertical air temperature gradient dT I dz such as: Qh = -C KdT I dz. When the air is stationary/the heat is transferred by molecular processes as a function of K, a molecular conductivity coefficient. However, since 1 0 1 the air is usually in motion and turbulent, the heat is conducted much more efficiently by air eddies and K is replaced by an eddy conductivity coefficient Ah. Then, uncertainties in Qh arise because of the lack of information on Ah, the latter being a property of air motions rather than a property of the air itself. The rate of heat transferred through the sea surface by conduction is generally positive dur ing the summer time (May to September) since the air temperature is usually greater than the water temperature. However, values are generally quite small and relatively unimportant compared with the other terms discussed above. From annual mean values (Pickard and Emery, 1982) Qh is near 0 W / m 2 over much of the B.C. coast, which also agrees with results from Tabata (1958). The conduction term can be more important in coastal areas, especially after a significant upwel l ing event has occurred. However, variations in Qh stil l remain small compared to those in Qs and Qb and may be neglected in most situations. 5.2.4 Evaporation/Condensation (Qe) This term is always negative over the ocean because heat is taken away from the water when evaporation occurs. The rate of heat loss is expressed by Qe = Fe- Lt where Fe is the rate of evaporation of water and L, is the latent heat of evaporation. Whi le Lt can be found in a table, Fe may be approximated by three different methods. The first of these is called the Pan Method and involves direct measurements of Fe. The latter are usually not very accurate because of the difficulties in reproducing the real atmospheric conditions and make the method not very useful. 1 0 2 A second method, called the f low method, is based on the fol lowing eddy diffusion-f low formula Fe = -Ae(df / dz) where Ae is the eddy diffusion coefficient for water vapour through the atmosphere and df I dz is the gradient of water vapour concentration (humidity) in the air overlying the sea surface. Because of difficulties similar to those mentioned in section 5.2.3, this formula is often replaced by a semi-empirical formula such as: Fe = lA(es —ea)-W where es is the saturated vapour over the sea-water and ea is the actual vapour pressure in the air measured at 10 m above sea level so that (es -ea) represents df I dz. W is the wind speed measured at 10 m height and gives an approximation of the variations of Ae. In the summer the SST is usually lower than the temperature of the overlying air, so that the latter is more stable. A s a result, the evaporation rate is reduced compared wi th winter values. From Tabata (1958), the rate of heat lost by evaporation in the vicinity of Triple Island is near -45 g -ca l / cm 2 / day (-21 W / m 2 ) between May and September and the annual mean value is near -89 g -ca l / cm 2 / day (-43 W / m 2 ) . Final ly, a third method for estimating Qe makes use of the heat budget equation. This method would not be val id for the present work because it requires the assumptions of no advection (Qv-0) and a steady state (QT=0) which are not true over the spatial and temporal scales involved in this thesis. A description of this method can be found in Pickard and Emery (1982). In general, variations in the rate of heat loss due to evaporation may be important for the daily SST fluctuations, although the values remain smaller than those associated wi th the incoming short-wave radiation. From the semi-empirical formula, variations in Fe, and therefore in Qe, are caused mainly by the variations in the w ind speed, which assist the evaporation process. The values for [es -ea) are 103 relatively small, especially during the summer because of generally small differences between the air temperature at 10 m height and the SST. 5.2.5 Advection (Qv) Besides the energy transfer processes described above, the movement of water, i.e. horizontal advection, upwel l ing and /o r mixing, may play an important role in affecting the SST on relatively short spatial and temporal scales. These advection terms are positive when there is inflow of warm water and negative when there is inf low of cold water, mixing and/or upwell ing. In particular, significant amount of heat may be advected by entrainment and wind-induced upwel l ing, especially in coastal areas. In most cases, values are not available due to the lack of measurements. However, an appreciation of the advective term can be done by reference wi th the other heat budget terms. 5.2.6 Total Heat Gain/Loss (QT) The total transfer of heat at the air-sea interface is estimated by the sum of the four heat flux terms across the surface of the ocean (Qs, Qb, Qh and Qe) plus the advection term (Q v). From a study by Tabata (1958), there is a net gain of heat into the sea near 200 g -ca l / cm 2 / day (97 W / m 2 ) from May to September and the year-to-year deviations are mainly due to modification of solar radiation by cloud cover for much of the summer season. 104 5.3 Ekman Layer Theory When the w ind blows over the ocean, a frictional force is exerted on the surface. In the atmospheric Ekman layer, the lower boundary condition for the momentum equations requires continuity of velocity and stress at the surface of the ocean, which creates a vertical w ind shear. The latter is responsible for a vertical exchange of momentum acting as a dr iv ing force for water movements and ocean currents. If u and w are the horizontal and vertical velocities, p the density, then the vertical flux of horizontal momentum is given by p u w. (per unit area). The mean value of this flux over a large area during a long period of time is equal to the mean stress T which is related to the w ind speed u by the fol lowing relationship: T = CDpu\u\ (5.1) For light w ind speeds, usually taken at 10m above the ocean's surface, there is a linear relationship between u and CD . However, at high winds, a correction is needed to account for high sea states. From Gi l l (1982), the drag coefficient is given as follows: j 1.1 x 10 3 for u < 6m / s CD=\ (5.2) I (0.61 + 0.063M) x 10"3 for 6m/ s < u < 22m / s Figure 26 shows the summer alongshore component of the w ind stress at the four buoy stations between 1990 and 1994. Negative w ind stress means that the w ind is b lowing from the northwest and is favorable for upwell ing along the west coast of north America according to the theory and previous studies (Al len, 1980, Ikeda et al, 1984 and others). From May to September, more than 70% of the w ind stresses are 105 less than 0.16 N / m 2 over Queen Charlotte Sound and off northern Vancouver Island. A long coastal areas, this percentage may increase to 90% (Manual Weather Hazards Manual , 1990). The fol lowing table (Table 20) gives simple statistics for the alongshore w ind stress components at the buoys for the study period. As mentioned in Chapter 4, the alongshore w ind is obtained by rotating the u and v-components by 45° . The summer mean alongshore w ind stresses are generally from the northwest, although mean values are relatively small. From min imum and maximum values, southeasterly w ind stresses reach generally higher strength than the northwesterly ones. Southeasterly w ind stresses are associated with frontal systems and stronger pressure gradients, while northwesterly w ind stresses accompany ridges wi th generally weaker pressure gradients. 106 1990: Buoy sites Max (N/m 2) Min (N/m 2) Mean (N/m 2) Std (N/m 2) 46204 0.15 -0.13 -0.01 0.04 46207 0.17 -0.14 -0.02 0.05 1991: Buoy sites Max (N/m 2) Min (N/m 2) Mean(N/m 2) Std (N/m 2) 46204 0.24 -0.14 0.00 0.05 46185 0.28 -0.19 0.00 0.06 1992: Buoy sites Max (N/m 2) Min (N/m 2) Mean (N/m 2) Std (N/m 2) 46204 0.39 -0.14 -0.01 0.06 46207 0.29 -0.15 -0.02 0.05 46208 0.24 -0.31 -0.02 0.08 40185 0.64 -0.25 -0.02 0.09 1993: Buoy sites Max (N/m 2) Min (N/m 2) Mean (N/m 2) Std (N/m 2) 46204 0.21 -0.15 -0.01 0.05 .46207 0.19 -0.16 -0.02 0.05 46208 0.10 ' -0.28 -0.04 0.07 40185 0.29 -0.28 -0.03 0.08 1994: Buoy sites Max (N/m 2) Min (N/m 2) Mean (N/m 2) Std (N/m 2) 46204 0.28 -0.11 0.01 0.05 46207 0.25 -0.11 0.00 0.05 46208 0.12 -0.19 0.00 0.05 40185 0.26 -0.15 0.00 0.06 Table 20. Maximum, minimum, mean and standard deviation for wind stress data during the summer time (Day 1 (May 1st) to day 153 (September 30th)). 1990 data are missing for the buoys 46185 and 46208, and 1991 data are missing for the buoys 46207 and 46208. Values are in N/m 2 -V • 107 There are two sources for the frictional force: one is the molecular friction which may be neglected in most aspects of the dynamics of ocean motions. The second force is related to the degree of instability of the motion: when the motion becomes turbulent, the non-linear terms in the Navier-Stokes momentum equations give rise to terms that have the physical character of friction. In most cases, friction forces achieved through turbulence are much more effective in re-distributing momentum and water properties. The horizontal stresses TX and T at the ocean's, surf ace form a vector representing the force per unit area applied on the sea surface. If the ocean is div ided into a series of thin layers, this force tends to set a motion to the top layer and thus exert a stress on the layer underneath, If an infinitesimal layer of thickness 8z is considered, the stress on the layer below can be.approximated by: [ix - (8zdxx I dz), ry - (Szdty I dz))-A n equal and opposite stress is exerted on the base of the original layer, and the net force per unit area on that layer is the difference between the stress on the top and bottom as follows: {dxx I dz,dty / dz)Sz The resulting force per unit mass that tends to accelerate the f luid is: \/p(dTx/dz,drJdz) and the vertical shear is measured by dxl dz where % = [xx,xy). 108 Ultimately, there is a depth where the momentum transmitted by the w ind vanishes and so do the frictional forces. Typically, the depth of the oceanic top boundary layer is 10m to 15m for calm summer days. Dur ing a study on w ind-driven currents over the Oregon continental shelf, A l len (1980) found a surface Ekman layer of about 20m depth under variable alongshore winds up to lOm/s . A s a result of the Coriol is force, the wind stress, applied at the ocean surface, forces a surface current to its right (Northern Hemisphere). The latter rotates gradually to the right wi th increasing depth down to the bottom of the surface Ekman layer. The net depth-integrated transport in the surface Ekman layer is directed at 90° to the right of the w ind stress direction. The Ekman transport leads to convergence/divergence of mass and hence, by continuity, to vertical motion of water into or out of the boundary layer called "Ekman pumping." In areas of convergence, a downward vertical motion pushes the isotherms down. In reverse cases, an upward vertical motion responds to divergence of surface waters which rises the isotherms, therefore, cools the upper mixed layer. The magnitude of vertical velocity wE just outside the surface Ekman layer can be found by integrating the continuity equation: du dv dw _ dx dy dz (5.3) 109 Integration is made wi th respect to z across the surface Ekman layer. Wi th the boundary condition of no vertical motion at the ocean's surface, i.e., w=0, the integral takes the fol lowing form: (5.4a) \udz + — vdz + w(0) - w(-h) = 0 (5.4b) volume fluxes which, by integration are equal to — and - — respectively, where p is the water density and / is the Coriolis parameter. It is a val id approximation to integrate the Ekman transport from (-oo to 0) instead of (-h to 0) because this transport is assumed to be zero below the level -h. Assuming that the variation of / is small compared to the variation of the wind stress and solving 5.4 for wE, the Ekman pumping velocity, gives: 5.3.1 Upwelling Theory Along the west coast of B.C., under northwesterly winds, the stress transmitted to the ocean's surface produces a net Ekman transport away from the coast, leading to a compensating onshore f low below the surface layer and an upward vertical motion (5.5) 1 1 0 near the coast called upwell ing. Such a process is observed by the appearance of relatively cold, dense water at the surface near the coastline and the transport of dissolved nutrients from greater depths to the euphotic zone. The offshore Ekman transport results in an area of low pressure along the coast relative to the pressure offshore. Then, the horizontal pressure gradient drives a current from the high towards the low pressure area (towards the shore) in the region just below the surface Ekman layer. Within this region, called the geostrophic interior, there is a balance between the pressure gradient and the Coriol is force which deflects the motion to the right (northern hemisphere). The resulting southerly f low below the surface Ekman layer is assumed to be geostrophic. The vertical structure of the ful l water column depth requires another Ekman layer below the geostrophic interior, with a thickness comparable to the surface Ekman layer. In the bottom Ekman layer, the ocean's floor exerts a drag on the current producing a mass flux at 135° to the left of the geostrophic current, so that the latter is deflected towards the coast. Then, by continuity, there is an upward vertical motion induced at the coast.' Upwel l ing is probably the main cause of short-term (two to ten days) SST fluctuations along the west coast of B.C. during the months of July and August. Typical values for the upwel l ing velocity are found between 1 to 10 m /day (Allen, 1980). The main part of the response is confined to a distance from the coast of the internal Rossby radius (typically 20 km along the coast of B.C.). Off the Oregon coast, the vertical motion has been observed over the continental shelf wi th in a distance of 10-15 km from the shore (Allen, 1980). Over much of Queen Charlotte Sound, upwel l ing is reduced and cold waters do not always reach the surface. A four month simulation run by a simple two-layer upwell ing model over Hecate Strait and 111 Queen Charlotte Sound has shown the 7°C isotherm rising from 200m to about 130m wi th the surface temperature simultaneously increasing from 7.8°C to 11.0°C due to surface heating (Ma, 1992). In such a situation, the SST, as measured by buoys may not be appropriate to detect temperature anomalies in fish marine environment below the surface. From (5.5), P / 1 ' an appreciation of upwel l ing responses can be obtained from the vertical displacement Az as follows: Assuming upwell ing favorable-winds in a two-dimensional Cartesian coordinate system (x, z) with the x-axis aligned across shore and positive onshore, and the z-axis vertical, the equation 5.4b leads to: — = -wa (5.6) Pf where w is the upwel l ing velocity and a is the Rossby radius. The alongshore variations of the current have been neglected since it is the cross-shelf mass balance that is of particular concern in upwell ing situations. The Rossby radius a is defined as follows: a = A h P f (5.7) 1 1 2 A n approximation of the Rossby radius over the study area can be obtained from survey data of August 23rd, 1950. The fol lowing values taken near Kains Island: A p / p = 2.24 x 10~3, / = 1.x 10"4 s' 1, the thickness of the layer H=78m and g = 9.8m / s2, correspond to a Rossby radius near 13 km. Just south of the buoy 46204, survey data give a Rossby radius near 20 km. From 5.6, w is given by: w « ^ - (5.8) By integrating the relation 5.8 on both sides wi th respect to time, we get the upward isotherm displacement due to upwell ing: Az=pfaj\t\dt (5.9) 5.3.2 Two-Layer Model The model described so far represents an ocean where the distributions of water properties such as density, temperature and salinity are continuous through the whole depth of the ocean. Al though this model is realistic, it is not computationally easy to solve. However, a solution can be found if the ocean is div ided into layers. In this model each layer has a constant density profile, different from one layer to the next so that the variation of the vertical structure of the water column is represented. A two-layer model wi th constant densities p, and p 2 representing the values in the mixed layer and the deep ocean is a good approximation for coastal regions where there is significant river runoff giving rise to an upper layer of low 1 1 3 salinity over a deep layer of much higher salinity, wi th a sharp halocline between them (Pond and Pickard, 1983). Al though this model contains only two vertical modes of oscillation; a barotropic mode and the first baroclinic mode, it sti l l remains a good approximation because in many cases involving long-period hydrostatic motions, most of the energy is usually contained in these first two modes (LeBlond and Mysak, 1978). The barotropic mode is one where the isobars and isopycnals are parallel so that the f low is uniform across the layer, while the first baroclinic mode presents tilted isobars compared wi th isopycnals so that the f luid in each layer flows in opposite direction. The latter case results in no horizontal net flux. In mathematical form, a two-layer model is expressed as follows: By reference to Figure 27, the upper layer has a density is p,, horizontal velocity components t^and vx, an equil ibrium depth //,.and a surface elevation z = r](x,y,t). Similarly, the second layer is characterized by p 2 , u2, v2, H2 and z = -Hx +h(x,y,t) where the total depth H is equal to Hx+H2. Approximations for the two-layer shallow water model: 1. For phenomena of relatively small scale (near 100 km), the Coriol is parameter, / = 2QsinO, may be assumed constant. The frame of reference used is called: f-plane. 2. When the horizontal velocities are large compared to the vertical velocities and the horizontal scales of motions are greater than the depth of the f lu id, then the pressure perturbation is independent of depth and there is a balance between the 114 vertical pressure gradient and the buoyancy force. This is called the hydrostatic approximat ion. 3. If the density variations are small, then their effect on the mass of the f luid can be neglected, but their effect on the weight must be retained. This is called the Boussinesq approximation Wi th these approximations, the linear shallow water equations wi th w ind forcing are: ^ L - y v l = - ^ + - ^ (5.10a) dt dx p,/Y, for the upper layer, and du2 . dr\ , dh /c 11 \ - T ^ - > 2 =-8w--8T~ ( 5 ' l l a ) at ox ox (hi r dri , dh / C H U N for the lower layer, h is the upward displacement of the interface and g' is the reduced gravity for a two layer f luid as follows: 8' = 8(P2-Pi)/Pi 115 The continuity equations for the upper and lower layers have the form: d(r] + Hx-h) | (dux | dvx dt \dx dy d{h) (du~, dv. • + H. = 0 dt 2 j wy2 dx dy = 0 (5.12a) (5.12b) The terms in brackets express the divergence of the horizontal transport. From 5.12a, the divergence of the f luid is causing a downward displacement of the free surface, whi le from 5.12b, it is resulting in an upward displacement of the interface between the two layers. For the baroclinic mode, the free surface displacement is smaller than the interface displacement so that the r igid l id approximation can be made. Then the equation of continuity 5.12a becomes: d(-h) dt du, dv, — L + — 1 dx dy = 0 (5.13) The difference in velocity between the two layers (also referred at the amplitude of the baroclinic mode) is obtained by subtracting 5.11 from 5.10 to give: du ^ , dh xr 3 - - ^ = ^ — + - f -dt dx pxtlx dv ^ ,dh T ¥ dt dy pxHx (5.14a) (5.14b) where u = ux-u2 and v = v, - v2 which can be thought of as the amplitude of the baroclinic mode of a two-layer system. 1 1 6 With the definit ion for u and v , the equation of continuity is given by: dh du dv _ dt dx dy (5.15) The equations 5.14 and 5.15 describe the baroclinic response. 5.3.2.1 Coastal Upwelling Using The Two-Layer Model Consider a particular case in which the surface Ekman transport is directed offshore wi th a vertical wal l at the coastline (i.e. at y=0). The vertical depth H and the w ind stress T x are constant, and T =0. In upwell ing situations, variations of the surface current in the alongshore direction can be neglected and d I dx —» 0. Then 5.14 and 5.15 are combined into a forced shallow-water equation wi th the fol lowing solution: f T V h = -f u = {P8' J V exp] 1— expj V v exp v f-y\\ V a j f-yW t V t V a JJ where c is the speed of long internal waves and is given by: (5.16a) (5.16b) (5.16c) c2 = JHXH2 (Hx + H2) and a=c/f is the Rossby radius (Gil l , 1982). (5.17) 117 Figure 28 illustrates this solution wi th the coastline at the eastern boundary and a w ind b lowing towards the equator. The surface offshore transport produces upwel l ing at the coast. From G i l l (1982), a w ind stress of 0.1 N / m 2 , g'=0.03 m / s 2 , H 1=100 m, and H 2 » H x result in an upwell ing velocity near 5 m/day . In the upper layer, Figure 28 shows a coastal jet which is formed by an increase of the alongshore current. In the lower layer, there is a poleward undercurrent f lowing in the opposite direction of the w ind (see G i l l , 1982 for more details). Whi le coastal upwel l ing decrease SST, G i l l (1982) mentioned that downwel l ing cannot warm the surface layer. Therefore, a succession of upwel l ing and downwel l ing events still tends to present cold waters along the coast. A n appreciation of the upwel l ing velocity can be obtained from the present SST data as follows: Consider at first a calm warm upper layer of temperature Ti and depth h over a cold lower layer of temperature TQ (sketch 1): 118 Without heat exchange, the thermocline would deepen due to mixing when the w ind start to blow. Then, the temperature of the upper layer becomes T2 and its depth increases to h + dh while the temperature of the lower layer remains the same (T0) (sketch 2): By conservation of heat, the fol lowing relationship applies (Tl-T0)h = (T2-T0){h + dh) (5.18) which gives the new sea surface temperature T2 due to w ind mixing only: 2 (h + Sh) (5.19) N o w , consider both upwel l ing and wind mixing at the same time. Then, the. thermocline wou ld move up due to upwell ing which results in divergence of the f low near the surface. The vertical displacement is given by 5.9. However, w ind mix ing wou ld move the thermocline back down towards its original level, as in the above sketch 2, and decrease the vertical displacement. The resulting upwel l ing 1 1 9 velocity can be obtained by differentiation of 5.19 with respect to time and by substituting d(8h)l dt = w. Therefore, dT2 ={T0-Tl)w dt h (5.20) as 8h 0 where w is calculated from the equation 5.8. 1 2 0 Figure 23. Distribution of the stratification parameter calculated for the spring tide (left) and the neap tide (right) (Jardine et al, 1993). 121 Figure 24. Daily short-wave radiation Q s in watts per square metre (w/m') received at the surface of the ocean in absence of cloud. From Pickard and Emery (1982). Figure 25. Long-wave radiation Qb in watts per square metre (w/m^) from a water surface as a function of surface temperature and the overlying relative humidity in absence of cloud. From Pickard and Emery (1982). 1 2 2 Wind Stress - Summer 1990 Wind Stress - Summer 1992 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Days (46204) Days (46204) 20 40 60 80 100 120 140 Wind Stress - Summer 1991 Days (46185) 20 40 60 80 100 120 140 Days (46185) Figure 26. Alongshore wind stress component in Newton per square metre (N/m^) at the buoy stations for the a) summer 1990 (Day 1 (May 1st) to day 153 (September 30th)), b) summer 1991, c) summer 1992. See Table 20 for missing data in 1990 and 1991. 123 Wind Stress - Summer 1993 Wind Stress - Summer 1994 -i ; ; ; ; ; ; I I ; ; ; ; ; i L _ 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Days (46207) Days (46207) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Days (46204) Days (46204) 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Days (46185) Days (46185) Days (46208) Days (46208) Figure 26. (continued), d) summer 1993 and e) summer 1994. 1 2 4 z -H * z - -nix, y, t) Upper layer density p\ Lower layer density pz Figure 27. Representation of two superposed shallow homogeneous layers of fluid. Hi , H 2 are the depths of the layers for a fluid at rest and H=Hi+H2 is the total depth. The z coordinate increases upward where z = Tj(x,y,t) is the surface elevation and z — —Ht + h(x,y,t) is the position of the interface between the layers. From Gill (1982). Rising I interface l< Rossby radius >j Figure 28. The solution for local upwelling at an eastern boundary. There is a coastal jet created in the upper layer in the same direction of the wind and an undercurrent in the opposite direction. From Gill (1982). 125 Chapter 6 Comparison Between SST, Synoptic Weather Maps and Wind Stress From the previous chapter, the main factors responsible for the warming or cooling of the surface of the ocean over the short time scale (a few days to a few weeks) in the study area are the fluctuations in the cloud cover and wind. The cloud cover directly affects the amount of incoming short-wave radiation reaching the surface of the ocean and the net amount of long-wave radiation leaving the water. The w ind is important for evaporation, upwel l ing (or downwell ing) and vertical mixing. The emphasis is put on upwel l ing for the present study. In this chapter three specific events are described from visual inspection in terms of SST variations and compared wi th synoptic weather maps, alongshore winds and /o r w ind stresses and cloud covers. Such analyses help to explain how the pressure patterns over the northeastern Pacific may interact wi th the upper layer of the ocean through the response of the surface water temperature variations over Queen Charlotte Sound and north of Vancouver Island. The three events, covering 30 to 32 days, have been selected based on an examination of the surface w ind , wi th an emphasis on the direction because of the importance of the induced Ekman transport and vertical motion on the SST (Ikeda and Emery, 1984, Jardine et al, 1993, Fang and Hsieh, 1993, Staples and Hsieh, 1994). The first two events, covering the periods from July 9th to August 8th (day 70 to day 100) of 1990 and from July 3rd to August 4th (day 64 to day 96) of 1992, are dominated by northwesterly winds. The third event, extending from June 29th to July 29th (day 60 to day 90) of 1993, shows variable winds. 126 To help understand the fol lowing analysis, SST, w ind stress and cloud cover time series are presented at the end of this chapter while synoptic weather maps can be found in appendices C , D and E. SST is in degrees Celsius and w ind stress is in N / m 2 wi th positive values for Southeasterlies and negative values for Northwesterlies. C loud cover is in oktas. In the fol lowing analysis, estimated values of the heat flux have been obtained from climatological data wi th Qs =+230 W / m 2 and Qb=-95 W / m 2 , both in absence of clouds, Qe at -25 W / m 2 and Qh near 0 W / m 2 . The resulting heat flux transferred to the ocean and available to increase the SST is nearly +115 W / m 2 in a cloud free environment during the periods selected. However, this value does not include any contributions from the advective term. The effect of clouds on daily AQS and AQb is considered by using the conversion factors seen in sections 5.2.1 and 5.2.2. The daily variations in latent and sensible heat fluxes (AQe, AQh), which depend mainly on changes in the w ind speed, are generally much smaller than those involved wi th the short and long wave radiation and have been neglected for simplification. 6.1 Event 1: July 9th to August 8th of 1990 (day 70 to day 100) 6.1.1 Overview Of Synoptic Weather Situation (Figures. CI to C31): At the beginning of the period, a weakening low pressure system was moving into the Gul f of A laska, forcing cold fronts to move across Queen Charlotte Sound on day 73 (July 12th) and on day 76 (July 15th). Thereafter, a ridge of high pressure dominated over the northeastern Pacific unti l day 93 (August 1st), whi le a trough of low pressure established itself along much of the B.C. coast from day 78 (July 17th) to 127 day 83 (July 22nd). After day 93 (August 1st) and until day 100 (August 8th) a series of frontal waves moved across the B.C. coast. 6.1.2 Comparison Between Synoptic Pressure Pattern, Wind, Cloud and SST Data (Figures. 29, 30, 31 and CI to C31): (Offshore data were available only for the buoys 46204 and 46207 for this event. See Figure 2 for locations.) Day 70 to day 77 duly 9th-16th): A t the beginning of the event, synoptic maps show south to south-southeasterly winds reaching 15 knots (7.8 m/s) over Queen Charlotte Sound on day 71 (July 10th) ahead of a cold front. Weaker winds (less than 10 knots or 5m/s) were observed around northern Vancouver Island due to the Cali fornian ridge extending over the island. A weak northwesterly w ind component began to develop over Queen Charlotte Sound on day 72 (July 11th) as a ridge of high pressure started to form near the Queen Charlotte Islands. In the wake of the cold front, the northwesterly w ind component gained some strength wi th values reaching just above 15 to 20 knots (7.8 to 10.3 m/s) over Queen Charlotte Sound on day 73 (July 12th). The northwesterly w ind speeds were 8.3 m / s and 8.7 m /s at the buoys 46207 and 46204 respectively just behind the front, which correspond to w ind stress values near 0.10 N / m 2 and 0.11 N / m 2 . Meanwhi le, the SST remained relatively constant over much of the area during the first 3 days (day 70 to day 73) except for a 0.6°C increase at Kains Island on day 73 (July 12th), which may be explained by a decrease of the cloud cover causing an increase in the heat flux into the water. Behind the front, the SST dropped by 0.02°C and 0.03°C at the buoys 46207 and 46204 respectively and by 0.7°C at Kains Island, suggesting an upwel l ing response over much of the study area. However, comparing the cooling of the surface waters at coastal (0.7°C) and offshore stations (0.02°C and 0.03°C) suggests that the upwell ing was more efficient at the coast. This 128 difference agrees wi th the equation 5.16a for the upwel l ing velocity showing an exponential decay away from the coast. Dur ing the fol lowing two days (days 74-75/July 13th-14th) a thermal trough developed over the southern interior of B.C. resulting in a northeasterly f low aloft near the coastal mountains and a lee trough along the west coast of Vancouver Island. A lso , there was ridge of high pressure off Vancouver Island extending northward across the Queen Charlotte Islands and into the northern interior of B.C. The resulting pressure gradient along the B.C. coast kept the Northwesterlies relatively constant over Queen Charlotte Sound with no significant variations in the SST unti l another front moved across Queen Charlotte Sound late on day 75 (July 14th). The cloud cover increased to above 6 oktas in the vicinity of this front, whi le the northwesterly w ind stress values increased slightly to near 0.10 N / m 2 behind it. This was reflected by a SST drop of 0.4°C at the buoy 46207 and 0.1°C at the buoy 46204 on day 76 (July 15th) and as much as 1.8°C at Kains Island on day 77 (July 16th). In more detail, the cloud cover increased from near 3 oktas to 6.5 oktas on day 76 (July 15th) at Cape Scott. This corresponds to a reduction in the amount of incoming solar radiation of approximately 35% and a reduction in the net loss of back radiation of approximately 40%. The estimated heat fluxes were +77 W / m 2 on day 75 (July 14th), down to +37 W / m 2 the next day. Despite the positive heat flux still transferred to the ocean on day 76 (July 15th), the SST did fall by 1.8°C during that day at Kains Island, which correlates with upwelling-favorable w ind stress and an increase in the cloud cover. W ind data are not available for Kains Island; however, by reference wi th upstream data, it is reasonable to estimate the upwelling-favorable w ind stress at nearly 0.10 N / m 2 as the second front moved 129 over northern Vancouver Island on day 76 (July 15th). The latter can certainly explain the SST variations via an upwel l ing response. From the relation 5.8, the upwel l ing velocity "w" at Kains Island was near 4 m/day on day 76 (July 15th), which gives a cooling of 0.4°C by using the relation 5.20. However, the observed cooling was 1.8°C suggesting other influences such as evaporation and /o r mixing. Advect ion of cold water is also possible, but cannot be confirmed from available data. A lso , the relation 5.20, which involves an upper layer wel l mixed wi th a uniform temperature "T" over a thermocline at a depth "h" , may justify the difference between the observed and calculated SST variation due to an over-estimation of the depth "h" of the upper layer. Further offshore, cloud cover values are not available from buoy data but it is reasonable to assume cloud cover values near 6 oktas as the front moved near the sites of these buoys. Then, similar heat budget arguments can be made for offshore waters, which indicates some upwel l ing occurring also over Queen Charlotte Sound, but to a lesser degree. The calculated upwel l ing velocity was near 4.3 m/day at the buoys, giving a cooling near 0.5°C. The latter over-estimates the real cooling at the buoy 46204 by approximately 0.4°C, but is relatively close to the value observed at the buoy 46207. Day 77 to day 83 duly 16th-22nd): Behind the second cold front, the ridge was weaker and the Northwesterlies were generally decreasing unti l day 79 Quly 18th) at the buoy 46207 and unti l day 80 (July 19th) at the buoy 46204. Meanwhile, the sky was clearing and the cloud cover was near 3 oktas at Cape Scott on day 77 and day 78 Quly 16th and 17th). Then, the estimated heat flux was +76 W / m 2 which can explain a general increase of the SST beginning on day 78 (July 17th), except one day earlier at the buoy 46207. Furthermore, a larger increase in SST was initiated on day 80 (July 19th) over much 130 of the area as the scattered clouds were dissipating. On day 81 (July 20th), there was only 1 okta of cloud reported at Cape Scott while no cloud was observed at Mclnnes Island. Noteworthy is a trough of low pressure wel l developed over much of the B.C. coast, especially on day 81 (July 20th), giving northwesterly w ind stress values near 0.09 N / m 2 at the buoy 46207, but somewhat weaker at the buoy 46204 (0.03 N / m 2 ). Despite the upwelling-favorable w ind stress, the SST was increasing, especially on day 80 (July 19th), showing a dominant effect from the incoming heat flux, unless there was an intrusion of warm water into the area. However, the SST at Kains Island began to decrease on day 82 (July 21st), which can be explained by coastal upwel l ing. This cooling phase also correlates wi th an increase in cloud cover to near 5 oktas at Cape Scott, despite an estimated heat flux near +50 W / m 2 . Further offshore, despite Northwesterlies, the buoy data show the warming process continuing unti l day 83 (July 22nd), suggesting the incoming, heat flux and /o r an intrusion of warm water into Queen Charlotte Sound being dominant. However, weaker warming of the surface waters near the buoy 46207, under stronger Northwesterlies, suggests some effect from upwell ing in that area. Other cooling mechanisms are also possible. Day 83 to day 87 (Tuly 22nd-26th): Between day 83 (July 22nd) and day 86 (July 25th), the Northwesterlies increased slightly, especially at the buoy 46204, as the trough was fi l l ing and the ridge bui ld ing offshore. A lso, fog and clouds invaded part of the area, especially along the coast, giv ing mainly obscured conditions near the northern end of Vancouver Island and along the mainland coast during this three day period. On day 83 (July 22nd), the estimated heat flux was +20 W / m 2 at Mclnnes Island and Cape Scott. Further 1 3 1 offshore, synoptic weather maps and cloud cover data for Cape St. James indicate more sunshine for an estimated heat flux near +65 W / m 2 . Meanwhi le, upwel l ing-favorable w ind stress were still occurring at the buoys 46204 and 46207 wi th values between 0.01 N / m 2 and 0.08 N / m 2 . Then, a cooling episode began at the buoy sites near or just after day 83 (July 22nd), which can be attributed to upwel l ing but also correlates wi th an increase in cloud cover from 5 to 8 oktas. A t the coast, surface waters were generally cooling, which may also be caused by upwel l ing and /o r reduced incoming solar radiation due to a relatively high cloud cover. This cooling episode ended on day 87 (July 26th) as a weak trough of low pressure moved across the Alaska panhandle. In the wake of this trough, more sunshine and weaker northwesterly w ind stresses did promote a warming trend over Queen Charlotte Sound. • , Day 88 to day 90 duly 27th-29th): Dur ing this period the ridge re-built off Vancouver Island, across the Queen Charlotte Islands and over the northern interior of B.C. while a lee trough developed just west of Vancouver Island. As a result, the pressure gradient increased over much of the study area during that period; this is reflected by the increasing upwelling-favorable w ind stress values at the buoy sites. However, the SST was increasing over much of the area, despite the upwelling-favorable w ind stress between 0.02 N / m 2 and 0.06 N / m 2 . The SST increased by 0.8°C at Kains Island and by 0.2°C at the buoy 46207 with an estimated heat flux near +88 W / m 2 , suggesting a dominant influence from the incoming heat flux. O n the other hand, the SST decreased slightly at the buoy 46204, (0 .1°C) , which may result from upwel l ing . 1 3 2 Day 90 to day 96 (Tuly 29th-August 4th): O n day 90 (July 29th), the northwesterly w ind reached relatively strong values giving northwesterly w ind stresses up to 0.13 N / m 2 , while the cloud cover increased to almost 8 oktas at Cape Scott for an estimated heat flux near +20 W / m 2 . The SST, dropping by 1.6°C at Kains Island on that day, suggests an upwel l ing response. The calculated upwel l ing velocity is near 5.5 m /day which under-estimates the real cooling by 1°C. Noteworthy is the correlation between the cloud cover increase and the beginning of the cooling phase. Further offshore the SST decreased only very slightly at the buoy 46204 and remained nearly stationary at the buoy 46207 indicating that the upwell ing did not reach the surface over Queen Charlotte Sound. By day 92 (July 31st), the northwesterly w ind component began to decrease wi th the approach of a deep low pressure system from the Pacific. The alongshore w ind shifted to the southeast on day 94 (August 2nd) wi th maximum speed near 4.9 m / s at buoy 46207 and 6.0 m / s at buoy 46204 on day 95 (August 3rd). The corresponding w ind stress values were near 0.03 N / m 2 at the buoy 46207 and 0.05 N / m 2 at the buoy 46204. Meanwhi le, the sky was mainly cloudy with between 5 and 8 oktas reported at Cape Scott, Mclnnes Island and Cape St.James for an estimated heat flux between +20 W / m 2 and +55 W / m 2 . Despite Southeasterlies and incoming heat into the surface waters, a cooling phase was occurring at Kains Island between day 92 Quly 31st) and day 95 (August 3rd). Noteworthy is a sharper cooling on day 95 (August 3rd) at Kains Island and the buoy 46204 which coincides wi th an increase in cloud cover from below 6 oktas to near 8 oktas on the previous day. Between day 92 Quly 31st) and day 95 (August 3rd), the SST data for the buoy 46207 indicates a warming phase occurring further offshore, which suggests a dominant effect from the 133 incoming heat flux and /o r the advective term by intrusion of warm water over Queen Charlotte Sound. Thereafter, a series of frontal waves moving over northeast Pacific kept a dominant southeasterly w ind component wi th speeds up to 5.9 m / s at the buoys unti l the end of the event. The corresponding wind stress values reached 0.05 N / m 2 . Meanwhi le the amount of cloud remained generally above 6 oktas for relatively low values of heat flux between +20 W / m 2 and +42 W / m 2 . This was reflected by a cooling phase after day 97 (August 5th). The evaporation and mixing processes were probably important in this case. 6.1.3 Summary Dur ing this event, the heat flux remained positive, even under cloudy skies. However, it appears that variations in cloud cover have some effect in regulating the SST variations. By looking at cloud cover data, this may be seen on days 76 (July 15th) and on day 90 (July 29th), where a cooling of the surface waters was coincident wi th a significant increase in the cloud cover. Such correlation can also be seen from a general cooling episode initiated on day 83 (July 22nd) when the sky became completely obscured with clouds. However, an upwell ing response is shown by looking at small differences in the SST variations between the sites compared wi th the w ind stress. For example, between day 77 (July 16th) and day 83 (July 22nd), the warming process was weaker at the buoy 46207 where the northwesterly w ind stress was in fact stronger, therefore more upwelling-favorable. It is noteworthy that the cooling effect of the evaporation process may help to further understand the small SST variations. 1 3 4 From the atmospheric pressure patterns, the wind stress was upwel l ing favorable behind cold fronts and wi th a ridge of high pressure across the Queen Charlotte Islands. Wi th cold fronts, the northwesterly w ind stress reached similar values at both buoys as the fronts moved through. The SST response was stronger at Kains Island and weaker offshore especially at the site of the buoy 46204. The mean SST was nearly 1.6°C cooler at the coast compared wi th those further offshore (Table 21). Such upwel l ing distribution results from an exponential decay of the response away from the coast. Wi th a ridge extending across the Queen Charlotte Islands, Northwesterlies tended to remain longer over the area, which could have been reflected in stronger upwel l ing event. This could not be confirmed from the data, especially because of interferences from other cooling mechanisms having. Table 21 just below gives the mean, maximum and min imum SST wi th the standard deviation for this event. Max imum upwelling-favorable w ind stresses developed wi th the presence of a ridge extending across the Queen Charlotte Islands and a lee trough. The strongest values were observed offshore, as indicated by the data of the buoy 46207, due to the position of the trough. In general, the troughs of low pressure tended to be centered near the coast so that the atmospheric pressure gradient was stronger offshore. Therefore, Northwesterlies tended to be stronger over offshore waters as compared wi th nearshore areas. Stations SST(mean). SST(max) SST(min) STD . , ' Kains Island 13.7 14.8 12.2 0.8 46204 15.5 1.6.6 14.6 0.5 46207 15.3 16.5 14.7 0.4 Table 21. Mean, maximum and minimum SST (°C) with the standard deviation during event 1. 1 3 5 6.2 Event 2: July 3rd to August 4th of 1992 (day 64 to day 96) 6.2.1 Overview Of Synoptic Weather Situation (Figures Dl to D33): At the beginning of the period, the northeastern Pacific was under the influence of a broad ridge of high pressure located near 140°W. However, after day 67 (July 6th) a series of frontal waves moved across the B.C. coast. O n day 77 (July 16th) the offshore ridge re-built in the wake of the last frontal wave, while a lee trough developed along much of the coast. Except for a weak front that moved across the B.C. coast on day 82 Quly 21st), a ridge remained anchored over northeast Pacific unti l day 84 (July 23rd). Thereafter, a low pressure system moved into the Gul f of Alaska, sending a series of frontal waves across the B.C. coast unti l day 90 Quly 29th). Then, a ridge re-built and remained over the northeast Pacific unti l day 96 (August 4th). 6.2.2 Comparison Between Synoptic Pressure Pattern, Wind, Cloud and SST Data (Figures 32, 33, 34 and D l to D33): Day 64 to day 67 fluty 3rd-6th): Dur ing this first period, the pressure gradient ahead of the ridge was relatively weak so that the up welling-favorable w ind was generally less than 5 m / s at the buoys, slightly stronger at buoy 46208. This corresponds to w ind stress values near 0.03 N / m 2 , except near 0.05 N / m 2 at the buoy 46208. The cloud cover was above 6 oktas at Cape Scott, Mclnnes Island and Cape St. James during those three days giving an estimated heat flux between +20 W / m 2 and +40 W / m 2 . Meanwhi le, the SST increased during the first day and then decreased from day 65 to day 67 Quly 4th-6th). ! 136 In particular, the SST increased by 2.3°C at Kains Island on the first day while there were 8 oktas of clouds (estimated heat flux near +20 W / m 2 ) wi th northwesterly w ind stress near 0.01 N / m 2 . The latter appears relatively weak and probably not sufficient to initiate an upwel l ing response, which would give the incoming heat flux more importance. However, the SST at Kains Island began to drop on day 66 (July 5th) when the northwesterly w ind stress reached near 0.04 N / m 2 at the buoy 46207 wi th no significant changes in the cloud cover. The next day, the cloud cover decreased to near 7 oktas for an estimated heat flux near +30 W / m 2 and the upwelling-favorable w ind stress increased further to above 0.04 N / m 2 . This was reflected in a 2.3°C temperature drop at Kains Island suggesting an upwel l ing response. Further offshore, the SST behaved in a similar fashion at the buoy 46207 but wi th smaller temperature variations. A t the other sites (buoys 46204, 46185 and 46208), there is insufficient evidence to give satisfactory explanations for the SST variat ions. Day 68 to day 77 duly 7th-16th): O n day 68 (July 7th), an upper trough moved eastward across Queen Charlotte Sound and northern Vancouver Island early in the day. In its wake, the offshore ridge gained more strength, as wel l as the upwelling-favorable winds. Max imum speeds between 6.2 m / s and 8.0 m / s were recorded at the buoy sites near day 69 (July 8th), which correspond to upwelling-favorable w ind stress values between 0.05 N / m 2 and 0.09 N / m 2 . A lso, the cloud cover decreased to near 6 oktas giving an estimated heat flux near +43 W / m 2 . Meanwhile, the SST increased at Kains Island and the buoy 46204, which may be caused by more heat coming into the sea surface and /o r by an intrusion of warm water around northern Vancouver Island. Further offshore, the surface water was slightly warming at buoy 46208. However, the SST 137 was dropping at the buoys 46207 and 46185, which correlates wi th upwel l ing-favorable w ind stress. The first of a series of frontal waves moved across the area late on day 70 (July 9th) increasing the cloud cover to above 6 oktas at Mclnnes Island and to near 8 oktas at Cape Scott and Cape St. James. This increase correlates with a short cooling episode at Kains Island on that day, although the estimated heat flux was stil l positive (near +20 W / m 2 ) . O n the other hand, stronger upwelling-favorable w ind stress and coastal upwel l ing can be assumed for a few hours in the wake of the wave and could also justify this cooling episode. Meanwhile, the SST was increasing at most buoys unti l day 71 (July 10th) and after, especially the buoy 46204, suggesting a dominant response to the incoming heat flux, while the upwelling-favorable w ind stress was below 0.05 N / m 2 and decreasing. However, further offshore the SST was decreasing at buoy 46207 during that period, which could be explained by somewhat stronger Northwesterlies (wind stress values between 0.05 N / m 2 and 0.06 N / m 2 ) in that area, making an upwel l ing response more likely. A lso, more evaporation due to stronger winds was probably contributing to the cooling phase. O n day 72 (July 11th), a second frontal wave moved across the area and the northwesterly winds dropped to near 0 m /s at most buoy sites. The sky was mainly cloudy as the wave moved through with above 6 oktas of cloud reported at Cape Scott, Mclnnes Island and Cape St James on that day. The estimated heat flux was between +20 W / m 2 and +43 W / m 2 . These factors: weak upwelling-favorable winds (if any) and incoming heat, supported a warming trend for all stations. After day 74 (July 13th), a weak ridge built temporarily along the coast of B.C. and the cloud cover began to decrease. The increasing heat flux into the sea surface supported the warming phase unti l day 76-77 (July 15th-16th). Meanwhile the upwelling-favorable 1 3 8 winds (if any) were generally weak unti l the last frontal wave moved through the area early on day 77 (July 16th). Just before this wave moved across the area, the estimated heat flux was between +30 W / m 2 and +56 W / m 2 while the northwesterly w ind stress (if any) was below 0.02 N / m 2 . Day 77 to day 84 (Tuly 16th-23rd): After day 77 (July 16th), the offshore ridge built up over the northeastern Pacific giving further clear skies to the area. The estimated heat flux increased to +110 W / m 2 between day 77 and day 79 (July 16th-18th). By day 79 (July 18th), the ridge had rotated clockwise and was aligned through the Queen Charlotte Islands and the northern interior of B.C. Further east, there was a thermal trough over the southern interior of B.C. resulting in a lee trough along the B.C. coast. The intensifying pressure gradient generated upwelling-favorable winds wi th maximum values reaching between 8.7 m / s and 12.0 m / s at the buoy sites near day 79 and day 81 (July 18th-20th). The corresponding wind stress values were between 0.12 N / m 2 and 0.25 N / m 2 . A lso, there was a low pressure system moving into the Gul f of Alaska on day 79 (July 18th), squeezing the offshore ridge between 1 3 5 ° W - 1 4 0 ° W so that the pressure gradient and associated northwesterlies were relatively strong over Queen Charlotte Sound and northern Vancouver Island. A s a result of upwel l ing-favorable w ind stresses, a cooling phase began near or just before day 77 (July 16th) at most stations despite mainly sunny skies. Noteworthy is a rapid cooling that occurred on day 80 (July 19th) at Kains Island (3.7°C) as the coastal trough reached its max imum development. The upwelling-favorable w ind stress values reached between 0.10 N / m 2 and 0.23 N / m 2 , suggesting an upwell ing response. The calculated upwel l ing velocity was near 5 m/day which under-estimates the real cooling by as much as 3°C. The estimated heat flux near +90 W / m 2 at Cape Scott, 139 makes the upwel l ing process more evident. However, at the buoy sites the temperature decrease was much smaller (less than 0.4°C) , which suggests an upwel l ing response much weaker offshore as predicted by equation 5.16a. The calculated upwel l ing velocity was near 5 m/day which over-estimates the cooling by as much as 94%. O n day 81 (July 20th), a weak front moved across the northern part of the B.C. coast, bringing clouds into the study area. In particular, the cloud cover increased to near 5 oktas at Cape Scott, and Mclnnes Island on day 81 (July 20th), decreasing the estimated heat flux to near +54 W / m 2 . A lso, the upwelling-favorable winds decreased in behind the front, especially in the northern part of Queen Charlotte Sound. However the northwesterly w ind stress remained relatively strong unti l day 84 (July 23rd), wi th values above 0.6 N / m 2 due to the offshore ridge being still relatively strong. Meanwhi le, the SST generally decreased suggesting a dominant upwel l ing response. Day 84 to day 90 duly 23rd-29th): After day 84 (July 23rd), a low pressure system moving into the Gul f of Alaska sent a series of frontal waves across the B.C. coast until day 90 (July 29th). A s a result, the upwelling-favorable winds stopped and the clouds moved in. The estimated heat f lux varied between +20 W / m 2 and +43 W / m 2 from day 84 to day 90 (July 23rd-29th) except up to +65 W / m 2 on day 86 (July 25th). Meanwhile, the SST generally increased between 0.5°C and 1.0 °C , which suggests a dominant influence of the incoming heat flux. It is noteworthy that the increase in SST began later (on day 86/July 25th) at Kains Island, which correlates wi th a relative min imum in the 140 cloud cover. Thereafter, the SST began to fall on or just after day 90 (July 29th) wi th increasing Northwesterlies. Day 90 to day 96 duly 29th-August 4th): O n day 90 Quly 29th) the last frontal wave moved across the area and the ridge built offshore. O n days 91 and 92 Quly 30th-31st), the ridge was coupled wi th a thermal trough over the southern interior of B.C. and Washington State. There also was a weak lee trough just west of Vancouver Island extending along the eastern boundary of Queen Charlotte Sound. This increased the pressure gradient and the Northwesterlies over much of the study area. The upwelling-favorable w ind stress reached between 0.07 N / m 2 and 0.17 N / m 2 on day 92 Quly 31st). The SST dropped by 1.1°C at Kains Island, 0.7°C at the buoy 46204 and 0.8°C at the buoy 46185 which suggests an upwel l ing response. Noteworthy is the clearing during these days which re-enforced a dominant upwell ing response: more heat was available, but the SST d id fall. The estimated heat flux reached near +110 W / m 2 on day 92 Quly 31st). After day 92 Quly 31st), some clouds moved over the study area, especially in the southern portion. A lso, the lee trough was relatively weak which is reflected by a decrease in the Northwesterlies for the next fol lowing two days. Meanwhi le, the SST generally increased which may be explained by an estimated heat flux between +40 W / m 2 and+65 W / m 2 . However the SST was temporarily fall ing at Kains Island on day 94 (August 2nd) suggesting an intrusion of cold water. By day 95 (August 3rd), the Northwesterlies generally increased due to the offshore ridge, especially in the southern part of Queen Charlotte Sound and around the northern end of Vancouver Island. Then the SST dropped at Kains Island and the southern most buoys (46204 and 46207) suggesting an upwell ing response. Further north, the 1 4 1 SST increased slightly at the buoys 46185 and 46208, which may be explained by sunny breaks. 6.2.3 Summary Dur ing this event, two sequences are worthy of attention. The first one occurred between day 72 and day 77 (July llth-16th) while the SST generally increased at all stations. Dur ing this short period, a series of frontal waves moved across the B.C. coast so that the alongshore w ind stress was not favorable for upwel l ing. A s a result, the SST was more vulnerable to the incoming heat flux across the surface of the ocean which increased from +20 W / m 2 on day 72 (July 11th) to near +100 W / m 2 on day 77 (July 16th). A lso, as each wave approached the area, the winds shifted temporarily to being from a more southerly direction, making advection of warm waters from the south possible. These wind variations are not shown by the daily w ind data since they may occur within less than 24 hours, but are inferred from work experience as a meteorologist. The second interesting sequence followed immediately upon the first one and shows a dominant effect from the advective term via the upwel l ing process. Between day 77 and day 84 (July 16th-23rd), the pressure pattern was dominated by a ridge of high pressure offshore and a trough along the B.C. coast giving Northwesterlies to Queen Charlotte Sound. In particular, the ridge extended across or just northwest of the Queen Charlotte Islands and into the northwestern interior of B.C. from day 79 to day 81 (Julyl8th-20th), generating relatively strong upwel l ing-favorable winds over Queen Charlotte Sound and in the vicinity of the northern end of Vancouver Island. This sequence was reflected in the SST by a general 142 decrease for several days at all stations. In particular, the SST dropped by 2.9°C at Kains Island, by 1.2°C and 1.6°C at the buoys 46204 and 46207, and by near 4.0°C at the buoys 46185 and 46208. Less cooling occurred offshore which is in agreement wi th the exponential form of the upwel l ing velocity (equation 5.16a). The upwel l ing-favorable w ind stress was stronger in the northern part of Queen Charlotte Sound, justifying a stronger cooling phase near the buoys 46185 and 46208. The fol lowing table (Table 22) gives the mean, maximum and min imum SST dur ing this event wi th the standard deviation. Stations SST(mean) SST(max) SST(min) STD Kains Island 13.3 15.9 11.8 1.0 46204 14.3 15.0 13.5 0.4 46207 14.5 15.2 13.5 0.4 46185 14.2 15.9 11.3 0.9 46208 13.2 14.4 9.8 0.9 Table 22. Mean, maximum and minimum SST (°C) with the standard deviation. 143 6.3 Event 3: June 29th to July 29th of 1993 (day 60 to day 90) 6.3.1 Overview Of Synoptic Weather Situation (Fig El to E31): Dur ing the first two days of this period (day 60-62/June 29th-July 1st) much of the B.C. coast was under the influence of a weak low pressure system. Then, a ridge started to develop offshore, especially in the wake of a weak frontal wave that moved across the B.C. coast on days 63-64 (July 2nd-3rd). This ridge remained anchored over the northeastern Pacific unti l day 81 (July 20th) and was coupled wi th a trough of low pressure along much of the B.C. coast between day 69 and day 77 (July 16th). Thereafter, a series of relatively weak frontal waves moved across the B.C. coast from day 82 to day 90 Quly 21st-29th). 6.3.2 Comparison Between Synoptic Pressure Pattern, Wind, Cloud and SST Data (Fig. 35, 36, 37 and E l to E31): (The following analysis does not include the area near the site of the buoy 46208 because cloud cover data were not available for Cape St.James during that period. ) Day 60 to day 64 (Tune 29th-Tuly 3rd): O n days 60 and 61 (June 29th-30th) a weakening low pressure system gave light and variable winds to Queen Charlotte Sound and the northern end of Vancouver Island. Meanwhi le the sky was mainly cloudy with an estimated heat flux between +20 W / m 2 and +40 W / m 2 . As a result, the SST was increasing at all stations. O n day 62 Quly 1st), the offshore ridge began to bui ld so that Northwesterlies started to develop over Queen Charlotte Sound and the northern end of Vancouver Island. Northwesterly winds between 2 and 5 m/s were reported at the buoy sites on days 62-63 Quly lst-2nd), corresponding to upwelling-favorable w ind stress values 144 between 0.01 N / m 2 and 0.04 N / m 2 . Meanwhile, the SST decreased at buoys 46204 and 46185, which reflects a dominant upwell ing response. The SST increased slightly at Kains Island and the buoy 46207, which can be explained by the heat flux (estimated at +40 W / m 2 ) being dominant. Between day 63 and day 64 (July 2nd-3rd), a weak frontal wave moved across Queen Charlotte Sound and Vancouver Island so that the upwelling-favorable winds stopped. Instead, there was a southeasterly wind component reported at most buoys wi th speeds near 3 m / s for w ind stress values near 0.02 N / m 2 . A lso, there were near 6.5 oktas of clouds reported at Cape Scott and Mclnnes Island on day 64 (July 3rd) for an estimated heat flux near +37 W / m 2 . Meanwhile, the SST increased slightly at most stations, which can be due to the heat flux into the surface waters and /o r an intrusion of warm water. Day 65 to day 82 duly 4th-21st): O n day 65 (July 4th), a broad high pressure system developed over the northeastern Pacific, as wel l as Northwesterlies over Queen Charlotte Sound and around Vancouver Island. A lso, clouds were dissipating so that the incoming heat flux was increasing from an initial value near +37 W / m 2 . On day 66 (July 5th), the heat flux was estimated between +55 W / m 2 to +70 W / m 2 . Meanwhile, a cooling phase had started at most stations near day 66-67 (July 5th-6th), showing clearly an upwel l ing response. The cooling phase started one day earlier at Kains Island, which suggests that upwel l ing is more efficient near the coast. Some cooling from evaporation and /o r mixing could also help starting this cooling phase. 145 Later, the offshore ridge continued to bui ld along a northwest-southeast line and reached its maximum development on day 68 (July 7th). As a result, the estimated heat flux reached maximum values between +60 W / m 2 in the south and +110 W / m 2 in the north. The upwelling-favorable w ind stress was also relatively strong wi th values near 0.14 N / m 2 in the south to 0.20 N / m 2 in the north on that day. This was reflected by further upwell ing at all stations. In particular, the SST dropped by 2.1°C at Kains Island, 1.7°C at the buoy 46204, 0.3°C at the buoy 46207 and as much as 3.1°C at the buoy 46185 between day 65 and day 68 (July 4th-7th). The calculated upwel l ing velocity, between 6 m/day and 8 m/day , generally approximates the observed cooling. However, the cooling is largely under-estimated at the buoy 46185 and over-estimated at the buoy 46207. Noteworthy is the estimated heat flux largely positive on day 68 (July 7th) which re-enforced a dominant upwel l ing response. On day 69 (July 8th) the offshore ridge started to migrate westward and a trough developed along the coast. Then, the upwel l ing-favorable winds began to decrease. On day 70 Quly 9th) the values decreased and were between 0.09 N / m 2 and 0.14 N / m 2 . Therefore, there was less upwel l ing and a warming phase began at most stations. Meanwhile the cloud cover was increasing, but the estimated heat flux remained positive wi th values near +30 W / m 2 at Cape Scott and near +92 W / m 2 at Mclnnes Island. This supported the warming process. After day 70 (July 9th), the trough started to fi l l along the coast and the pressure gradient continued to weaken over Queen Charlotte Sound and around Vancouver Island. Then, the Northwesterlies continued to decrease unti l day 79 (July 18th). A lso , more clouds invaded the area, giving an estimated heat flux between +20 W / m 2 and +40 W / m 2 . Meanwhile, the warming trend continued at the stations located close to the shore, especially at the buoys 46204 and 46185. Noteworthy are short cooling episodes at Kains Island between day 73 and day 75 (July 12th-14th) and 146 on day 78 (July 17th) embedded in a general warming phase. These cooling episodes are correlated wi th the passage of two weak fronts, which suggests an upwel l ing response from a band of stronger Northwesterlies in the wake of each front. Part of the cooling may also be caused by the evaporation and/or mixing processes. Further offshore, the SST oscillated at the buoy 46207 wi th temperature variations generally less than 0.5°C unti l the warming phase started on day 76 (July 15th). From day 79 Quly 18th) to day 81 (July 20th), the offshore ridge and the Northwesterlies decreased further. The northwesterly w ind stress was between 0.01 N / m 2 and 0.02 N / m 2 , except near 0.04 N / m 2 further offshore (buoy 46207). Meanwhi le, the sky remained mainly cloudy giving an estimated net heat flux between +20 W / m 2 and +40 W / m 2 at Cape Scott and Mclnnes Island. However, the SST was generally fall ing at most stations suggesting a dominant effect from the advective term, most l ikely through an intrusion of cold water since the upwel l ing-favorable w ind stress was relatively small. Day 82 to day 86 duly 21st-25th): O n day 82 (July 21st), a deepening low pressure system moving from the west forced the ridge onto the coast. Then, clouds began to dissipate. The estimated heat flux near +20 W / m 2 on day 82 (July 21st) increased to near +30 W / m 2 on day 83 (July 22nd). The alongshore winds shifted from northwesterly to southeasterly on day 82 (July 21st) and were relatively strong. The southeasterly w ind stress was up to 0.16 N / m 2 dur ing those two days. Meanwhile, the SST decreased at all stations, especially on day 83 (July 22nd) when the wind speed was relatively strong. This 147 suggests a dominant effect from wind mixing, although a part of the cooling may be attributed to the evaporation process. O n day 84 (July 23rd), the offshore ridge re-developed in the wake of the low pressure system. As a result, the sky continued to clear and the upwelling-favorable winds re-developed over Queen Charlotte Sound and near Vancouver Island. Max imum heat flux near +70 W / m 2 occurred on day 85 (July 24th). Max imum upwelling-favorable winds were recorded at the buoy sites on day 86 (July 25th) when a lee trough developed just west of Vancouver Island. The corresponding w ind stress values ranged between 0.07 N / m 2 and 0.11 N / m 2 from the northwest. Meanwhi le, the SST was generally increasing, suggesting a dominant effect from the incoming heat flux. Day 87 to day 90 duly 26th-29th): O n day 87 (July 26th) a cold front approached Queen Charlotte Sound and the northern end of Vancouver Island spreading clouds to the area. The cloud cover increased to near 8 oktas on that day, which correlates with the onset of a cooling phase at Kains Island. From synoptic weather maps, winds were generally light and variable near the front. The northwesterly w ind stress was below 0.01 N / m 2 at the buoy sites and probably not strong enough to initiate an upwel l ing response. In fact, the SST increased at all buoy sites. After day 87 (July 26th), the cloud cover decreased to between 4 and 7 oktas, which supported the warming process offshore almost unti l day 90 (July 29th). Meanwhile, the SST continued to decrease at Kains Island unti l day 89 (July 28th) and then increased. 148 6.3.3 Summary The most important sequence during this period occurred between day 65 and 68 (July 4th-7th) while a ridge was bui lding just off the coast of B.C. The relatively strong Northwesterlies ahead of the ridge were reflected by a cooling phase at all stations due to upwel l ing. Noteworthy is the temperature drop being larger near the coast, especially at the site of the buoy 46185. Later, when a trough developed along the B.C. coast, the upwelling-favorable w ind stress continued, but at decreasing strength. This was reflected by a general warming phase of the surface waters, showing a direct response to the wind stress. Finally, a temperature drop near 0.7°C recorded at the buoy sites on day 83 (July 22nd) reflected a dominant effect from vertical mix ing due to southeasterly winds up to 25 knots (12.8 m/s) for w ind stress values near 0.25 N / m 2 . Some cooling from the evaporation process was also probably occurring. Table 23 gives the mean, maximum and min imum SST dur ing this event wi th the standard deviation. Stations SST(mean) SST(max) SST(min) STD Kains Island 12.7 14.9 11.1 0.9 46204 14.0 14.9 12.8 0.5 46207 13.5 14.7 12.9 0.4 46185 13.6 14.7 11.0 1.1 Table 23. Mean, maximum and minimum SSr r (°C) with the standard deviation. 149 Figure 29. Wind stress time series for event 1: July 9th-August 8th (day 70-100) of 1990. Wind stress values are in Newton per square metre (N/m2). 150 i 1 1 1 1 1 1 1 1 1 1 i r Days (Kains Island) T 1 1 1 1 1 1 1 1 1 I I I r J i I l I I I i i l I I I 1 1 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 Days (46204) 1 1 1 1 1 1 1 1 1 1 1 I I T _i i i i i i i i i i 1 1 1 1 1 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 Days (46207) Figure 30. SST time series for event 1: July 9th-August 8th (day 70-100) of 1990. SST values are in degree Celsius (°C). The dotted lines correspond to the summer mean SST. 151 O1 1 1 1 1 1 I I I I I I L 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 Days (Cape Scott) 0' — 1 1 1 1 1 1 1 1 1 I i i i i I 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 Days (Cape St.James) Figure 31. Cloud cover time series for event 1: July 9th-August 8th (day 70-100) of 1990. Cloud cover values are in oktas. 1 5 2 T 1 1 1 1 1 1 1 1 1 1 1 1 1 r —i 1 1 1 1 1 i i i i ' ' ' i l l 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 Days (46204) T I I I I I 1 1 1 I 1 1 1 1 T - I 1 1 1 1 I I ' ' i i i i i i 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 Days (46207) T i i i i 1 1 1 1 1 1 1 1 1 r - I 1 1 1 1 1 1 I I 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 Days (46208) Figure 32. Wind stress time series for event 2: July 3rd-August 4th (day 64-96) of 1992. Wind stress values are in Newton per square metre (N/m2). 153 Days (46208) Figure 33. SST time series for event 2: July 3rd-August 4th (day 64-96) of 1992. SST values are in degree Celsius (°C). The dotted lines correspond to the summer mean SST. 154 \ 0 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 Days (Cape Scott) Days (Cape St.James) Figure 34. Cloud cover time series for event 2: July 3rd-August 4th (day 64-96) of 1992. Cloud cover values are in oktas. 1 5 5 T 1 1 1 1 1 1 1 r Days (46185) Figure 35. Wind stress time series for event 3: June 29th-July 29th (day 60-90) of 1993. Wind stress values are in Newton per square metre (N/m^). 156 Figure 36. SST time series for event 3: June 29th-July 29th (day 60-90) of 1993. SST values are in degree Celsius (°C). The dotted lines correspond to the summer mean SST. 157 Figure 37. Cloud cover time series for event 3: June 29th-July 29th (day 60-90) of 1993. Cloud cover values are in oktas. 158 Chapter 7 Satellite Remote Sensing of Sea Surface Temperature The previous chapters have explored the relationship between buoy data and information from lighthouse stations. By looking at specific events in terms of SST, cloud cover and w ind stress, a relatively good correlation was found between Kains Island and the buoy sites, although the daily SST tends to vary more at Kains Island than at the buoy sites due to coastal effects. Now, to go a step further in this study, it wou ld be interesting to f ind out how representative of the whole study area are those buoys in terms of SST. This is done in this chapter by using Advanced Very H igh Resolution Radiometer (AVHRR) satellite imagery. The latter w i l l give a larger picture of the SST distribution over the study area and w i l l al low a further understanding of the relationship between coastal and buoy stations. However, it was not possible to get a time series for any events analyzed in the previous chapter because of frequent fog and clouds over Queen Charlotte Sound and the northern end of Vancouver Island. Nevertheless, useful information could be obtained from a few A V H R R infrared satellite images by visual inspection of the radiance distribution and by plotting transects. Analyses of these transects are presented in section 7.7. 7.1 Data Source Satellite images have been provided by the advanced very high resolution radiometer (AVHRR) on board the N O A A 10 and N O A A 11 satellites. These 159 satellites are in sun-synchronous orbits 870 km above the earth so that a fixed geographical location on the earth is always observed at the same local solar time on each pass every day. This can be achieved by a suitable choice of the orbit's inclination (near 100°) so that the orbit plane is fixed relative to the. sun as the earth travels around the sun once a year. For the N O A A satellites, a complete orbit takes approximately 103 minutes so that they are repeated 14 times a day. A s the earth rotates underneath, the satellites scan from north to south over one face of the earth and south to north over the other face giving a sampling rate of two images a day for a given location. The A V H R R sensor is an optical instrument, providing multi-spectral imaging by sensing reflected solar radiation in two visual or near visual wavebands (channels 1 and 2) and by sensing thermal emissions in three infrared wavebands (channels 3,4 and 5) as follows: Band 1 0.58- 0.68 urn (Visible), Band 2 0.73- 1.10 urn (Near infrared), Band 3 3.55- 3.93 urn (Thermal infrared), Band 4 10.3- 11.3 | im (Thermal infrared), Band 5 11.5- 12.5 urn (Thermal infrared). The main waveband used in this chapter corresponds to Band 4 in the infrared (IR) region. This is the best channel to measure the SST because of a relatively good balance between the peak of the thermal emission and the atmospheric transparency in the 10.3 -11.3 urn wavelength range. On the other hand, Band 1 is used for cloud detection and land references. The spatial resolution of A V H R R imagery is 1.1 km at the sub-satellite point and decreases to near 7 km at the edge of the swath that is 2580 km wide. The calibration 160 of IR images is relatively important to get the most accurate absolute values of the SST. It is achieved wi th the use of a blackbody cavity at a known temperature carried on board the satellite. The radiometer resolution is measured by a noise equivalent difference temperature (NEAT) parameter where A T refers to the apparent change in temperature. For typical SST wi th values between 270 and 300 ° K , the resolution is near 0.13 - 0.10 ° K for Bands 4 and 5. In practice, the value of N E A T is generally higher due to calibration errors, atmospheric effects (absorption, re-emission, scattering and reflection). Large absolute errors may also be introduced by sub-pixel (i.e. less than 1.1 km) cloud contamination (Robinson, 1994). To remove these errors, further calibration has been done in this thesis by using in situ SST measurements from buoys. A summary of the basic principle underlying the operation of IR sensors to get SST values is presented in the fol lowing section. 7.2 Characteristics Of Infrared Radiation Infrared radiation is emitted by the sea surface as a result of vibrational and rotational excitations within the molecules of water. The spectral range of infrared radiation includes Band 3, 4 and 5. Based on Planck's theory and the emission of a perfect emitter at 300 ° K (average temperature of the sea surface), the peak of the ocean radiation spectrum is at 10-12 um (Fig. 38). The spectral emittance for a body at temperature T is described by Planck's relationship as follows: M W ' n * * S , X T ) - V 1 & i W / m 2 A U n i Z 1 ) 161 where A is the wavelength (um), Cy and C 2 are constants wi th the fol lowing values: C, =3.74 10"16 W / m 2 and C2=1.44 lO- 2 m°K. M(A) is the emittance (also called exitance) for a particular A which expresses the radiant flux density per unit bandwidth centered at A , leaving a unit area, regardless of the direction. Equation 7.1, based on ideal thermodynamic laws, is val id only if the source of emission is a perfect emitter (i.e. blackbody). For the ocean surface, which is not a perfect blackbody, the emittance is described in term of its spectral emissivity, e(k). The latter is the ratio of the emittance from a real body at temperature, T, compared to a blackbody at the same temperature, T, and wavelength, X: The spectral emissivity is almost constant wi th respect wi th temperature, but varies significantly wi th wavelength. e(A) = M(A) M(A), BBT RT (7.2) The integration of equation 7.1 over all wavelengths results in the total exitance of a blackbody, M , which is the Stephan-Boltzman Law as follows: M = oT4 W / m 2 (7.3) where s = 5.669 x 10"8 W / m 2 ° K 4 is Stephan's constant. The wavelength, A,, max, corresponding to the peak emission is given by Wien's displacement law: A m a x r = 2897 (um°K) (7.4) 1 6 2 and is shown in Figure 38 where the peak of the emission for a surface at 300 ° K such as the ocean occurs near 10 jtm. The Wien's displacement also shows that by making use of sensors detecting the range covered by Band 4 (10.5 -11.3 urn), it is possible to measure sea surface temperature. 7.3 Satellite Measurements The A V H R R is a sensor measuring the radiant flux, (j), in watts (w) per unit sol id angle, Q,, in steradian (sr), per unit normal area (m) per unit bandwidth (urn). By definit ion, the radiant flux per unit solid angle leaving a point source in a specific direction is the radiant intensity: / = _ w / s r (7.5) wi th the assumption that the sea surface is Lambertian, i.e. emits isotropically (Robinson, 1985). If the sea surface is divided into small areas, the radiant intensity leaving a unit area normal to the direction of the remote sensing is called the radiance, L, and is given by the following: L = w / s r / m 2 ' (7.6) J(Acosc9) 163 By using the equation 7.1, since the emitted radiation is independent of the direction, the radiance for a given wavelength is given by: L(A) = ^ ^ W /m2/ Sr/um (7.7) K In other words, L(A) is the radiant flux, O , in watts (w) per unit solid angle, Q , in steradian (sr) per unit normal area (m) per unit bandwidth (mm) for a blackbody. Then, it is possible to f ind the temperature, T, by using (7.7) and inverting (7.1). This temperature, T, is the brightness temperature for a particular wavelength assuming that the source is a perfect emitter. If the emissivity is known, then the true temperature can be found from the radiance of the blackbody using the fol lowing expression: M(X)BB = K h { X ) — < d 10-6 W / m 2 / u m (7.8) e(A) For the sea surface, the emissivity is approximately 0.98. It varies little wi th temperature, wavelength and surface roughness. However, the value may be affected by surface slicks or debris (Robinson, 1985). The A V H R R sensor measures L{X), and the apparent temperature of the sea surface is obtained by using equations 7.8 and 7.1. However, the IR radiation measured by the satellite, L(A) , is emitted by a very thin layer of molecules at the surface of the ocean. For the wavelengths of channel 4 and 5 (10.5 to 12.5um), the thickness of that layer is less than 0.1 m m (Grassl, 1976). Therefore the satellite remote sensing derived sea surface temperature is a skin temperature. Typical values of the 164 temperature deviation between the skin temperature and the bulk sea surface temperature measured by buoys and lighthouse stations within the top metre are near 0.01 ° K to 0.05 ° K wi th the skin of the ocean usually being cooler than the layer just underneath (Schluessel, 1990). Useful interpretation of SST measured by the A V H R R can be made if the thermal vertical structure of the near surface layer is known. The next section presents an overview of a typical upper layer thermal structure, fol lowed by some limitations it may have on the satellite remote sensing of SST. 7.4 Upper Layer Thermal Structure O n the basis of Planck's theory, the sun emits radiation like a blackbody at 6000 °K. The spectral irradiance reaching the sea level covers mainly the visible region and a small part of the infrared. The full spectrum falls approximately wi th in 0.30 urn to 2.5 urn wi th the peak emission near 0.50 (am (Fig. 39). Most of this energy is absorbed by the ocean within the first few metres (Fig. 39) and used to heat the surface layer or to promote photosynthesis (Thurman, 1994). The first wavelengths to be absorbed are the shorter ones, i.e., the infrared, followed by the red whi le the blue light can been seen down to 100 m due to scattering by water molecules and particles. A t 100 m, only 1% of the energy incident on the surface remains. The incoming solar energy is transmitted into the ocean by conduction and turbulent mix ing due to winds and waves. As mentioned in section 5.2.3, conduction is a very slow process and allows only a fraction of the heat to be transferred downwards. However, turbulent mixing is much more efficient in redistributing the energy downward. Therefore, this process controls the 1 6 5 temperature of the mixed layer depth which can be as thick as 200-300 m a t mid-latitudes in the open oceans but as little as 10 m in protected coastal areas dur ing the summer. Below the mixed layer, the temperature decreases rapidly to approximately 1000 m. This layer forms the permanent thermocline. Then, from 1000 m to the bottom of the ocean, the temperature decreases, only very slowly to 0°C to 3°C (Fig. 40). Dur ing the summer, due to an increase of the incoming solar radiation and generally less turbulent mixing, a seasonal thermocline usually forms above the permanent one. It is not unusual to observe a seasonal thermocline in May near 20 m with a mixed layer above it. As the summer progresses, the surface mixed layer deepens so that seasonal thermoclines reach near 40 m in September (Fig. 41). The actual depth of the summer upper mixed layer varies wi th time and space as it is closely related to local w ind forcing (section 5.3.2) and solar radiation. Furthermore, dur ing warm sunny days without much wind stress, a diurnal thermocline can also develop on top of the seasonal one. Such a thermocline, occurring generally at a few metres deep wi th a typical temperature difference near 1-2°C (Open series, 1991), is more localized and is destroyed at night due to heat losses to the atmosphere (Robinson, 1985) (Fig. 42). O n the other hand, the higher the surface temperature, the more stable is the upper layer due to the non-linear relationship between the density p and temperature T: A change in density Ap is larger for a unit change in temperature dT under warm conditions than cool ones. Hence, warm waters tend to be more stable than colder waters in similar environments. 1 6 6 7.5 Limitations Of IR Sensing Of The Ocean Surface The first and main limitation of satellite-sensed SST depends on the difference between the sea skin temperature and the bulk sea surface temperature because oceanographers are more interested in the latter when studying ocean dynamics and water properties. 7.5.1 Characteristics Of The Skin Layer The skin layer has a thickness always less than 1 mm (Grassl, 1976) which depends on the vertical heat flux through the air-sea interface. Its temperature is typically a few tenths of a degree Kelv in lower than the temperature measured just a few centimetres below (between 0.1 ° K and 0.5 °K) (Schluessel, 1990) as a result of the long-wave radiation emitted from the upper few micrometres of the ocean. The transfer of latent and sensible heat between the ocean and the atmosphere generally produces further cooling of the skin layer. The relative importance of these fluxes depends on the air-sea temperature and water vapor mixing ratio differences. It is also related to the surface wind speed which creates turbulent exchanges of heat and momentum. Only exceptionally the sea skin temperature was found to be higher than the near surface temperature just below. The sharpest temperature gradient wi thin the skin layer normally persists at w ind speeds up to 10 m /s , but it is destroyed by breaking waves at w ind speeds above 10 m/s . However, studies have shown that it takes only 10 to 12 s for the skin layer to redevelop after the w ind has decreased below the 10 m / s l imit (Schluessel, 1990). Differences between the surface skin temperature and the bulk temperature measured approximately 1 metre below the surface are found between -1.0 ° K to 1.0 ° K (Robinson, 1985, Schluessel, 1990). Correction for the skin effect is done in this thesis by making comparisons between the bulk temperatures measured from buoys with radiance values from A V H R R after the images have been processed. This method is the most common because of its simplicity (Schluessel, 1990). On the other hand, a study by Tabata (1981) off Vancouver Island showed good agreement between A V H R R measurements of the SST and the in situ SST measurements from ships. The fol lowing section gives details on the image processing. 7.6 Image Processing The first requirement when using A V H R R satellite data is that images be relatively cloud-free so that ocean dynamic process can be observed. Therefore, the images have been checked at the Institute of Ocean Sciences and/or the University of Brit ish Columbia for cloud contamination and only those wi th a low cloud content (generally less than 20%) have been kept. Then, available images wi th in the events of chapter 6 have been further checked for cloud contamination over the study area by visual inspection of Band 1 images and by using threshold values on enhanced Band 4 images. From 11 images, 3 have been selected for further processing: One image for July 31st (day 92), 1992 during event 2 and two others for July 7th and 8th (day 68, 67), 1993 during event 3. Each image was taken during the afternoon pass so that the SST measurements can be compared. 168 Further processing was performed on the images by applying a low-pass filter to block the high spatial frequency details. The 3 x 3 weighted-filter used is the fo l lowing: 1 1 1 1 4 1 1 1 1 This type of filter is called an "unequal weighted-filter" and was developed by Wang et al. (1983). Whi le a simple smoothing operation blurs the images, especially at the edges, this type of filter helps to reduce blurring. Then, the fol lowing high-frequency filter developed by Pratt in 1978 was applied to the images to accentuate or sharpen edges: - 1 - 1 - 1 -1 9 -1 -1 -1 -1 The image processing has been done by using the N I H Image program for the Macintosh. Further information about processing and /o r this program can be found in the user's guide provided by Macintosh. 169 7.7 AVHRR Observations The next fol lowing three sub-sections (7.7.1, 7.7.2 and 7.7.3) present the descriptions of satellite images. Each sub-section begins with a synoptic situation describing the sea-level pressure pattern over the northeastern Pacific and western Nor th Amer ica wi th in approximately 1 hour of the time the images was taken (a). This is fol lowed by a general description of clouds and winds from the synoptic weather map and /o r the satellite image (b). Also included are daily averaged in situ SST wi th anomalies (SSTA) and w ind stress from the buoy data for the day the image was taken. Then, the image description follows in "c", before an analysis of transects in "d" . These transects, taken from the images, show the spatial SST distribution between the stations. There is also a table in sub-section "d " including the mean, max imum and min imum calibrated SST with the standard deviation for each transect. Unless otherwise mentioned, SST refer to calibrated SST values in this section. The calibration of the radiance values has been made with in situ buoy data. 1 7 0 7.7.1 Image 1 The first image (Nll-19845) shown on Figure 43 was taken on day 92 (July 31st) of 1992 at 22:19 Pacific daylight time (PDT). a) Synoptic Situation: The sea-level pressure pattern over northeastern Pacific at 23:00 PDT on day 92 (July 31st) was dominated by a ridge of high pressure near 135-140°W (Fig. 44). Further east, there was a thermal trough over the interior of British Columbia and Washington State. b) Clouds, Winds and in situ SST: The sky was generally sunny, except for a band of cirrus, approximately 55 km wide, extending southeastward from north of Calvert Island (Fig. 43). A lso, there was a narrow band (near 10 km wide) of low clouds from southern Calvert Island extending into Queen Charlotte Strait. Northwesterly winds near 10 knots (5 m/s) were reported over much of the study area except for northerly 20 knots (10.3 m/s) at Cape Scott (Fig. 44). The buoy data indicate upwelling-favorable w ind stress near 0.07 N / m 2 in the South and up to 0.17 N / m 2 in the North, while the averaged daily in situ SST were generally below normal (Table 24). 171 Stations Daily SST (°C) Daily SSTA (°C) Wind Stress (N/m2) Kains Isld. 11.8 -1.9 46204 13.7 -0.5 -0.07 46207 14.8 +0.1 -0.07 46185 14.2 -0.5 -0.17 46208 13.7 -0.1 -0.17 Table 24. Daily in situ SST and SSTA in degree Celsius (°C), and wind stress (N/m2) for day 92 (July 31st) of 1992. c) Image Description (Fig. 43): Over the cloud-free area, the radiance distribution shows many variations that reflect a complicated flow pattern over the study area as described in Crawford et al. (1995). In particular, there is evidence of cold upwelled-water (deep blue color) just off Aristazabal Island and further north along the mainland coast, which has also been observed by Jardine et al. (1993). Then, a cold water plume (between 4 to 8 km wide) is seen approximately 20 km east of the buoy 46185 and southward between Midd le and Goose Island banks. Further south, the signature of the plume is hidden by the cirrus clouds. The SST is below 12°C along that plume and as low as 10.6°C just off Aristazabal Island. Then, there is a large (near 70 km in diameter) area of warmer water (light blue/white color) over Midd le bank wi th SST up to 15.7°C. This feature has been found to be relatively persistent in time by Jardine et al. (1993) and was called the "Moresby Eddy". To the west of this eddy, the image suggests a f low of cooler water (SST near 14°C) into the Moresby Trough. Then, 10 km southeast of Cape St. James, there is a small cold core (5 km in diameter) wi th SST near 11.9°C. Another cold area (SST near 12.4°C) appears 15 km southwest of Cape St. James. 172 Close to the mainland coast and south of Aristazabal Island, there is no evidence of cold upwelled-water and temperatures reach near 15.5°C just off Calvert Island. Warmer waters appear just north of Vancouver Island wi th values up to 16.6°C near Cook bank and east of it. A n area of cold water (deep blue and purple) is evident off Cape Scott wi th SST as low as 10.7°C. This feature was found to be the result of strong tidal mixing (Jardine et al. 1993). Finally, a large warm eddy (near 45 km in diameter) is shown 25 km west of Kains Island wi th SST up to 16.3°C. d) SST Distribution: 1. Transect #1: Buoys 46185 - 46204 (Figure 45a) (The transect extends from 10 km north of the buoy 46185 to 6 km south of the buoy 46204.) The SST is near 13.5°C at the initial point (0 km) where the satellite image suggests upwel l ing. It increases to 14.1°C near the site of the buoy 46185 and to near 15.7°C at the northeastern edge of the Moresby eddy. Then, the SST gradually decreases to just below 12°C in the cold plume before increasing again. Further south, pixels wi th radiance values above 155 (i.e., SST < 11°C) have been ignored because of cloud contamination. The satellite image suggests that cold upwelled-waters appear mainly north and east of the buoy 46185 and that the SST at the site of the buoy is more representative of those over the Moresby eddy. Further south, warmer SST is evident near the site of the buoy 46204 suggesting stratified conditions. The SST at the site of the buoy 46204 is near 15.2°C. 173 2. Transect #2: From the buoy 46207 to 46204 (Fig. 45b) (The transect extends from the site of the buoy 46207 to 7 km past the buoy 46204.) The SST distribution is relatively uniform between the two buoy sites. Moreover, the SST is the same at the sites of the two buoys (15.2°C). However, there is a band (18 km wide) wi th higher SST 20 km from the site of the buoy 46207. A lso, there is a band (25 km wide) of cooler water approximately 20 km from the site of the buoy 46204 that looks like tidally induced upwell ing. This phenomena has been observed in previous studies (Jardine et al, 1993). 3. Transect #3: From Kains Island to the buoy 46207 (Figure 45c) (The transect begins 2.5 km south of Kains Island and ends at the site of the buoy 46207.) The SST distribution within the first 105 km from Kains Island reflects cold waters from coastal upwel l ing and tidal mixing with temperature near 13.9°C just off Kains Island in Quatsino Sound and as low as 11.2°C 76 km from there. However, there is a band of warmer waters between distances of 33 km and 40 km from Kains Island which may be associated wi th a frontal area between the coastal upwel l ing and tidal mix ing area. O n the other hand, it may reflect the water temperatures of a warm eddy as mentioned in c). Further offshore (beyond 100 km from Kains Island), the SST reflect the water temperatures around the site of the buoy 46207 where the mixing seems weaker. Values increase to near 15.2°C at the site of that buoy, which is more than 3°C warmer than the waters just off Kains Island. Transects Mean SST(°C) Max. SST(°C) Min. (°C) STD (°C) 46185-46204 15.8 46207-46204 15.2 15.9 14.4 0.4 Kains I.-46207 13.8 15.7 11.2 1.1 Table 25. Standard deviation and mean, maximum and minimum SST (°C) along the transects. 1 7 4 7.7.2 Image 2 This image ( N l 1-24662) shown on Figure 46 was taken on day 68 (July 7th) of 1993 at 23:57. a) Synoptic Situation: O n day 68 (July 7th) of 1993 at 01:00 PDT, the pressure pattern over the northeastern Pacific was dominated by a strong high pressure system centered near 5 0 ° N / 1 4 6 ° W (Fig. 47). Further east, there was a thermal low over southeastern Brit ish Columbia. b) Clouds, Winds and in situ SST: The sky was generally sunny, except for traces of thin clouds far off Cape Scott and a small area of low cloud half way between Cape St. James and the mainland coast (Fig. 46). The pressure gradient was relatively strong over Queen Charlotte Sound and near the northern end of Vancouver Island wi th northwesterly winds up to 25 knots (12.8 m/s) corresponding to upwelling-favorable w ind stress up to 0.25 N / m 2 (Fig. 47). The daily in situ SST were below normal at all stations (Table 26). Stations Daily SST (°C) Daily SSTA (°C) Wind Stress (N/m2) Kains Isld. 11.4 -2.2 46204 13.6 -0.7 -0.15 46207 13.1 -0.9 -0.14 46185 11.0 -2.9 -0.20 46208 -0.23 Table 26. Daily in situ SST and SSTA in degree Celsius (°C), and wind stress (N/m2) for day 68 (July 7th) of 1993. 175 c) Image Description (Fig. 46): The radiance distribution is relatively uniform over much of Queen Charlotte Sound wi th a mean SST near 13-14°C. However, there are significant variations from offshore to coastal areas. In particular, there is clear evidence of upwel l ing (deep blue and purple colors) along much of the mainland coast, especially near Aristazabal Island and northward where coldest waters are found. A cold water plume extends offshore from Aristazabal Island as far as the site of the buoy 46185 and 60 km to the South. The SST near the site of that buoy is near 12°C but decreases to near 10.6°C within 15 km to the Northeast. The warm Moresby eddy shows again over Midd le and Goose Island banks, but covers a larger area (up to 100 km wide). The highest SST (near 14.2°C) appears over Goose Island bank, approximately 60 km off the northern end of Aristazabal Island. Further west, there is a sharp temperature gradient (1.7°C within 3.5 km) near the 1000 m depth contour. A lso , a large cold eddy (50 km in diameter) is evident approximately 55 km south of Cape St. James wi th SST near 11°C. Over the southeastern area, cold upwelled waters extend from the coast to approximately 20 km offshore and southward. There is a cold plume (SST near 11°C) extending from Calvert Island to approximately 35 km east of the buoy 46204. Co ld water is also evident off Cape Scott and Kains Island (purple and deep blue) where the SST is near 2-3°C cooler than the surrounding waters. Co ld waters extend wi th in approximately 35 km off Kains Island due to upwel l ing and up to 80 km off Cape Scott due to tidal mixing (Jardine et al, 1993). Further offshore, a large area of warmer waters (near 14°C) is evident beyond 50 km off Kains Island where a small eddy (10 km in diameter) can be seen. 176 cO SST Distribution: 1. Transect #1: From the buoy 46185 to 46204 (Fig. 48a). The SST is near 12°C at the initial point (10 km north of the buoy 46185) and near 12.3°C at the buoy 46185. The satellite image shows clearly cold-upwelled waters reaching the site of that buoy. Then, the SST decreases to a min imum of 11.2 °C towards the centre of the plume (approx. 10 km southeast of the buoy 46185). Further south, the SST increases gradually to reach a maximum value (14.0°C) over Goose Island bank (approx. 120 km southeast of the buoy 46185) and then decreases slightly towards the site of the buoy 46204. From the satellite image, it seems that cold upwel led water near the mainland coast flows southward into the Goose Island trough and towards the site of the buoy 46204. However, the SST at the site of the buoy 46204 is warmer (13.5°C) , suggesting less or no influence from upwel l ing. 2. Transect #2: From the buoy 46207 to 46204 (Fig. 48b) A s for Image 1, the SST distribution is relatively uniform between the two sites of these buoys suggesting similar dynamics and/or heating process along the transect. However, the SST is slightly lower near the site of the buoy 46204 which can be attributed to an intrusion of cold upwelled waters from the mainland coast. The SST is 13.7°C at the site of the buoy 46207 and 13.4°C at the site of the buoy 46204. 3. Transect #3: From Kains Island to the buoy 46207 (Fig. 48c) 177 ( Relatively cold upwelled-waters are indicated within the first 76 km from Kains Island. The coldest waters (near 10°C) are found within approximately 15 km of the coast. Then, the SST increases to near 12°C. Beyond 76 km, the SST increased by 1.7°C. A n area of cold waters (near 12.8°C) appears between a distance of 78 km to 125 km due to tidal mixing. Then, the SST rises to reach 13.7°C near the site of the buoy 46207 where water conditions are more stratified. Transects Mean SST(°C) Max. SST(°C) Min. (°C) STD (°C) 46185-46204 12.9 14.0 11.2 0.8 46207-46204 13.5 13.9 13.0 0.2 Kains I.-46207 12.5 13.8 10.3 0.7 Table 27. Standard deviation and mean, maximum and rrunirnum SST (°C) along the transects for day 68 (July 7th) of 1993 at 23:57 PDT. 178 7.7.3 Image 3 This image ( N l 1-24676) shown on Figure 49 was taken on day 69 (July 8th) of 1993 at 23:45. a) Synoptic Situation: O n day 69 (July 8th) of 1993 at 01:00 PDT, the high pressure centre had moved further to the northwest from the previous day and was located near 5 4 ° N / 1 4 9 ° W (Fig. 50). Further east, there was a cold front from the just east of the Queen Charlotte Islands extending to southeastern British Columbia. A lso, there was a weak lee trough just along the west coast of Vancouver Island. b) Clouds, Winds and SST: The sky was generally sunny over the study area, except for a few traces of cirrus clouds (Fig. 49). A lso, there was an area of low clouds around the northern end of Vancouver Island. The winds were generally from the northwest wi th speeds near 20 knots (10.3 m/s) (Fig. 50). The daily in situ SST were still below normal at most stations. 179 Stations Daily SST (°C) Daily SSTA (°C) Wind Stress (N/m2) Kains Isld. 11.1 -2.5 46204 12.9 -1.5 -0.09 46207 13.5 -0.6 -0.15 46185 11.2 -2.8 -0.14 46208 -0.19 Table 28. Daily in situ SST and SSTA in degree Celsius (°C), and wind stress (N/m2) for day 69 (July 8th) of 1993. c) Image Description (Fig. 49): The radiance distribution is stil l relatively uniform over Midd le and Goose Island banks wi th SST comparable to those of the previous day (13-14°C) . The cold upwelled-water reaching the site of the buoy 46185 has not progressed much further offshore, whi le its temperature has remained relatively constant off Aristazabal Island and northward. However, the cold water plume is now reaching approximately 10 km further south than the previous day. The core of that plume flows 20 km east of the buoy 46185 wi th SST values between 10.4°C off Aristazabal Island and 12.4°C further south. The highest SST (near 14.5°C) stil l appears over Goose Island bank. Further west, the sharp temperature gradient near the 1000 m depth and the cold eddy south of Cape St. James do not show significant change from the previous day. Over the eastern shore of Queen Charlotte Sound, the satellite image shows a. larger upwel l ing area than the previous day with a few filaments of cold waters extending southward from the mainland coast. Near Calvert Island, the upwel l ing area has expanded by approximately 5 km further west with the coldest waters (11°C) just north of the Island. A lso, the cold upwelled-water is reaching much further south 180 with its intrusion near the buoy 46204.evident. The SST is 12°C 9.5 km northeast of that buoy and 13.5°C at the site. Further south, there is an area of low clouds around the northern end of Vancouver Island hiding the SST distribution just off Kains Island. The cold waters resulting from tidal mixing are still visible off Cape Scott and are now spreading towards the buoy 46207. In fact, the SST has decreased by 1°C at the buoy from the previous day. Further south and approximately 50-60 km off Kains Island, an area of warm water is still evident, but the SST has decreased by near 1°C from the previous day. d) SST Distribution: 1. Transect #1: Buoy 46185-46204 (Fig. 51a) The SST is just below 12°C at the initial point (10 km north of the buoy 46185) and near 12.2°C at the buoy 46185, which is similar to the previous day. Further south, the SST is relatively low (between 11.5°C and 12.4°C) in the cold plume. Noteworthy is a sharp increase 70 km south of the buoy 46185 which is the southern l imit of that plume. Then, the SST oscillate around 13.8°C over a distance of approximately 60 km. The warmest SST (14.4°C) is reached at a distance of 115 km from the buoy 46185. Further south, the SST decreases again towards the buoy 46204 due to cold upwelled-waters reaching the area. 2. Transect #2: Buoy 46207-46204 (Fig. 51b) Once again, the SST. distribution is relatively uniform between the two buoys. However, the SST northeast of the buoy 46204 are decreasing due to the intrusion of 181 cold upwelled-waters towards that buoy as mentioned in c). Lower SST are also seen at the buoy 46207 due to tidally mixed waters spreading towards that buoy. 3. Transect #3: Kains Island-Buoy 46207 (Fig. 51c) The SST have been ignored for the first 45 km from Kains Island to eliminate any cloud contamination. The main feature is what appears as a front at a distance of 66 km from Kains Island. Noteworthy is that front now being 10 km closer to Kains Island compared wi th the previous day. A n area of relatively cooler waters appears between distances of 88 km and 116 km from Kains Island, which is associated wi th the core of mix ing area due to the action of the tide. Then, the SST slowly increases towards the buoy 46207. Transects Mean SST(°C) Max. SST(°C) Min. (°C) STD (°C) 46185-46204 12.5 13.9 11.2 0.8 46207-46204 13.2 13.6 12.5 0.2 Kains I.-46207 12.5 13.6 11.3 0.5 Table 29. Standard deviation and mean, maximum and minimum SST (°C) along the transects for day 69 0uly 8th) of 1993 at 23:45 PDT. 182 7.7.4 General Observations In Upwelling Conditions The satellite images show clearly that wi th upwelling-favorable winds the SST at Kains Island does not reflect those over much of Queen Charlotte Sound. Upwel l ing is seen near the northern end of Vancouver Island wi th SST near 2 °C colder than that 20 km to 30 km offshore. Further west, the transects between Kains Island and the buoy 46207 show a consistent temperature change at distances of 65 km to 75 km from Kains Island which is the limit of the mixing area. Then, beyond approximately 80 km off Kains Island, the SST increases and is generally a few degrees warmer at the site of the buoy 46207. The latter represents the conditions over a large part of Queen Charlotte Sound. North of Vancouver Island, the SST is relatively constant between the buoys 46207 and 46204, however, it generally decreases past the buoy 46204. There are a few cold water plumes extending within approximately 20 km from the mainland coast, which may occasionally reach as far as the site of the buoy 46204. Further north, the SST are relatively uniform beyond the upwel l ing area, wi th maximum SST generally reached over Goose Island bank, 30 km to 40 km north of the buoy 46204. However, the upwel l ing covers a much wider area wi th cold water plumes spreading up to 70 km off the mainland coast. In particular, upwel l ing seems a dominant feature near the site of the buoy 46185 where the SST is similar to that near Kains Island, but can be more than 1 °C cooler than that near the buoy 46204. Finally, the information is l imited near the site of the buoy 46208 due to cloud cover. However, upwell ing is seen wi th in approximately 25 km off the west coast of the Queen Charlotte Islands, but does not seem to have much influence on the water properties near the site of the buoy 46208. 183 i o 8 UJ o z WAVELENGTH, >jm Figure 38. Emission spectra at three different temperatures. From Robinson (1985). visioie wavelength (nm) Figure 39. Energy-wavelength spectrum of solar radiation at the surface of the ocean and at different depths (nm=nanometre=10"9m). (From Brown et al, 1989). 1 8 4 temperature (°C) ,0 5 10 15 20 mid- latitudes Figure 40. Typical mean temperature profile for mid-latitudes in the open oceans. temperature (°C) 8 10 s ~~~ September 100 u Figure 41. Temperature profiles showing successively the growth (solid lines) and decay (broken lines) of a seasonal thermocline in the Northern hemisphere. 1 8 5 Figure 42. Typical near-surface temperature profiles showing the diurnal thermocline during calm, slightly-mixed and night-time conditions. (From Robinson, 1985). 186 187 188 Figure 45. Satellite sensed sea surface temperature (°C) distribution for day 92 (July 31st), 1992 at 22:19 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km). 1 8 9 190 191 Figure 48. Satellite sensed sea surface temperature (°C) distribution for day 68 fluly 7th), 1993 at 23:57 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) 2.5 km south of Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km). 1 9 2 1 9 3 1 9 4 Figure 51. Satellite sensed sea surface temperature (°C) distribution for day 69 (July 8th), 1993 at 23:45 PDT from a) 10 km north of the buoy 46185 to 6 km south of the buoy 46204, b) the buoy 46207 to 7 km east-northeast the buoy 46204 and c) 2.5 km south of Kains Island to the buoy 46207. The star signs marked the location of the sites along the transects. Distances are in kilometre (km). 195 Chapter 8 Conclusion 8.1 Summary and Conclusions Summer daily SST over Queen Charlotte Sound and near the northern end of Vancouver Island have been studied using buoy and lighthouse data. From a visual inspection, the most evident cycle of SST anomalies time series occurs at 2 to 3 days due to the influence of weather systems. The statistical analysis does not show much similarity between the stations, especially between coastal and offshore stations. The correlation analysis indicates that most r-values are not significantly different from zero. However, it shows a significant correlation between the buoys 46204 and 46207 in 1992 and 1993. On the other hand, the w ind time series of the buoy stations are all significantly correlated wi th r-values above 0.60. Final ly, the spectral analysis shows generally more energy at low frequencies. In general, the statistical analysis has not been very successful because the time series present many uncorrelated fluctuations due to local effects. In chapter 6, an event-by-event analysis is done mainly by visual inspection: by looking at SST, windstress and cloud cover data along wi th synoptic weather maps. This analysis shows that the main cause of SST variations dur ing the summer is upwel l ing associated wi th a high pressure centre offshore, a ridge extending in the vicinity of the Queen Charlotte Islands and/or further north and a lee trough along the west coast of Vancouver Island. The Northwesterlies and upwel l ing response can be enhanced over Queen Charlotte Sound with the approach of a low pressure system from the west. In other situations, upwell ing can be triggered behind a cold front. However, relatively strong Northwesterlies do not persist over a long period of time and so neither does the upwell ing. The induced vertical velocity is near 4 m / d a y to 5 m / d a y during the peak of the events. However, the response is stronger near the coast, i.e. at Kains Island and the buoy 46185 and decreases at the offshore buoys as predicted by the theory. A second cause of SST variations is the changes in the incoming heat flux due to fluctuations in the cloud cover. The influence of clouds can be observed mainly when the upwelling-favorable winds are relatively weak or non-existent. However, the lack of consistency between cloud cover and SST time series does not permit conclusions wi th certainty. The,variations in latent and sensible heat fluxes have been neglected in the estimated heat flux compared wi th those involved in the incoming and back radiation. However, it appears that the daily variations of the latent heat flux from the evaporation process could help to further explain the SST variations, especially when upwel l ing is not dominant. Finally, a cooling effect due to vertical mixing could be occasionally detected wi th strong south to southeasterly winds which are hot frequent dur ing the summer t ime. In general, the SST is lower at Kains Island compared wi th the offshore buoys due to a stronger upwel l ing response. A visual inspection of the SST time series for specific events shows that the SST variations among the stations agree relatively wel l , especially between the buoys 46204 and 46207. On the other hand, the SST at Kains Island shows more short term variations, while that at the buoy 46185 manifests episodic strong cooling. The A V H R R satellite images confirm the difference between coastal and offshore SST during upwelling-favorable situations in which case the SST at Kains Island does not represent that over much of Queen Charlotte Sound. In fact, the most representative station of that area is the buoy 46207 which does not appear to be influenced much by upwell ing. Then, it wou ld be 197 interesting to compare the fish migration route and the northern diversion wi th the SST at that buoy. However, the fish may respond to factors other than SST! 8.2 Future Work In order to further investigate the SST variations, a more complete heat budget should be completed by looking at the air temperature. This would allow us to consider the evaporation term which may be relatively important dur ing upwel l ing events. A lso , a more accurate heat budget may explain the SST variations when upwelling-favorable winds are weak or non-existent. O n the other hand, atmospheric pressure data from the buoys could be used to make an index related to the northern diversion. 198 Bibliography Al len , J.S., Models of wind-dr iven currents on the continental shelf. A n n . Rev. F lu id . Mech. 1980. 12:389-433. Atk inson, L.P. and Blanton, J.O., 1986. Baroclinic Processes on Continental Shelves, Processes that affect stratification in shelf waters. Christopher N.K. Mooers, Series Editor, 130 pp., 1986. Brown, J. et al, Open Series: Seawater: Its composition, properties and behaviour, Pergamon Press, 165 pages, 1989. Crawford, W.R. Physical Oceanography of the Waters around the Queen Charlotte Islands. Submitted for publication in Ecology of marine and shoreline birds of the Queen Charlotte Islands, Brit ish Columbia. Crawford, W.R., Woodward, M.J., Foreman, M.G.G. , Thomson, R.E. Oceanographic features of Hecate strait and Queen Charlotte Sound in summer. Atmosphere-Ocean 33(4), 639-681. 1995. Emery, W.J. and Hamil ton, K. Atmospheric forcing of interannual variabil i ty in the Northeast Pacific Ocean: Connection with E l Nino. Journal of Geophysical research, vol . 90, no c l , pp 857-868, January 20th, 1985. Environment Canada. Marine Weather Hazards Manual . Ministry of supply and services, 1990. Catalogue No. En 56-74/1990E, ISBN 0-662-18193-X. Fang, W., Hsieh, W.W. Summer sea surface temperature variabil i ty off Vancouver Island from satellite data. Journal of Geophysical Research, Vo l . 98. No . C8: 14391-14400. 1993. Favorite, F. Surface temperature and salinity off the Washington and Brit ish Columbia coasts, August 1958 and 1959. Journal Fishery Research Board of Canada, 18(3), 1961. Foreman, M.G.G. , Henry, R.F., Walters, R.A., and Ballantyne, V . A . A finite element model for tides and resonance along the north coast of British Columbia. Journal of Geophysical Research, submitted. 1992. Freeland, H.J., Crawford, W.R. and Thomson, R.E. Currents along the Pacific coast of Canada. Atmosphere-Ocean, 22 (2), 151-172. 1984. 199 Garrett, C.J.R., Keeley, J.R. and Greenberg, J.R. Tidal M ix ing versus Thermal Stratification in the Bay of Funday and Gulf of Maine. Atmosphere-Ocean 16(4) 1978,403-423. 1979. . • ." , G i l l , A . E . Atmosphere-Ocean Dynamics. Academic Press, Inc. 662 pp., 1982. Grassl, H . The dependence of the measured cool skin of the ocean on w ind stress and total heat flux. Boundary-Layer Meteorology 10(1976) 465-474. 1976. Groot G . and Qu inn T.P. Homing migration of sockeye salmon, Oncorhynchus Nerka, to the Fraser River. Fishery Bulletin: Vo l . 85, No . 3: 455-469. 1987. V Hami l ton, K. A study of the variability of the return migration route of Fraser River sockeye salmon (Oncorhynchus nerka) Can. J. Zool . 63: 1930-1943. 1985 Herl inveaux, R.H. O n tidal currents and properties of the sea water along the Brit ish Columbia coast. Fisheries Research Board of Canada. Progress Reports N o 108, September 1957. Ikeda, M . and Emery, W.J. A continental shelf upwel l ing event off Vancouver Island as revealed by satellite infrared imagery. Journal of Marine Research, 42, 303-317,1984. Jardihe, I.D., K.A. Thomson, M .G. Foreman and P.H. LeBlond. Remote sensing of coastal sea surface features off northern British Columbia. Remote Sensing of the Environment, 45:73-84. 1993 Jensen, J.R. Introductory Digital Image Processing - A Remote Sensing Perspective, Prentice Ha l l , N e w Jersey, 385 pp., 1986. Kendrew, W . G . and Kerr, D. Climate of British Columbia and the Yukon Territory. Edmond Cloutier, C .M.G. , O.A., D.S.P.,Ottawa, 1955. 222 p. 1955.. Large, W.G. , Crawford, G.B. Observations and simulations of upper ocean response to w ind events during the Ocean Storms Experiment. Accepted by JPO for special Ocean Storms issue. 1994. Leblond, P .H. ,Dyck ,K. , Perry K., and Cumming, D. Runoff and precipitation time series for the coasts of British Columbia and Washington. Department of Oceanography, U B C , Manuscript Report no. 39, May 1983. Leblond, P.H. , Mysak, L.A., Waves in the Ocean, Elsevier, Amsterdam, 602 pp., 1978. M a , Hela i . A simpli f ied model of interannual water temperature variations in Hecate strait and Queen Charlotte Sound. Master's thesis. Department of Oceanography, U B C , August 1992. 2 0 0 Mysak, L.A. E l N ino , interannual variability and fisheries in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43: 464-497. 1986. Pickard, G.L. and Emery, W.J., Descriptive Physical Oceanography (4th edition), Pergamon Press, New York, 249 pp., 1983. Pond, S., Pickard G.L., Introductory Dynamical Oceanography, Pergamon Press, N e w York, 329 pp., 1983. Robinson, I.S., Satellite Oceanography: A n introduction for oceanographers and remote-sensing scientists, Ell is Horwood, England, 455 pp., 1985 Royer, T. O n the effect of precipitation and runoff on coastal,circulation in the Gul f of Alaska. Journal of Physical Oceanography, 9, 555-563. 1979. Schluesses, P., Emery, W.J., Grassl, H . and Mammen, T. O n the Bulk-skin temperature difference and its impact on satellite remote sensing of sea surface temperature. Journal of Geophysical Research, vol. 95, N O . C8, pp 13,341-13,356, August 15th, 1990. Staples, G.C. and Hsieh, W.W. Satellite A V H R R observation of a summer upwel l ing event off Vancouver Island. Submitted to Journal of Geophysical Research(Oceans). October 1994. Tabata, S. Heat Budget of the Water in the Vicinity of Triple Island, Brit ish Columbia. Journal of Fisheries Research Board of Canada, 15(3), pp. 429-451, 1958. Tabata, S. On the accuracy of satellite-observed sea surface temperatures, in Oceanography from Space, edited by J.F.R. Gower, 145-157, Plenum Publ ishing, New York, 1981. Thurman, H.V., Introductory Oceanography, 7th edition, McConn in R.A., N e w York, 550 pp., 1994. Webster, I., Farmer,D. Analysis of salinity and temperature records taken at three lightouse stations on the B.C. coast. Pacific Marine Science Report 76-11. March 1976. Xie, L. and Hsieh, W.W. Predicting the Return Migrat ion Routes of the Fraser Sockeye Salmon (Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. vo l 46:1287-1292. 1989. 2 0 1 Appendix A Decoding of Synoptic Report The model used on weather synoptic maps is the following: cH TT St? PPP W w w f N J ppa T d T d WR, C L N h RR h where ppp is the atmospheric pressure, ppa the pressure tendancy within the last 3 hours, wRt is the past weather conditions, RR is the precipitation in tenth of millimetre, C L / N H and h refer to the type, coverage and height of low clouds, Td Td is the temperature of the dew point, W and ww are the visibility and atmospheric conditions, TT is the air temperature measured 10 m above the ground and N is the cumulated amount of cloud (in oktas). For ship and/or buoy reports, the SST is indicated within brakets and below the temperature of the dew point Td. 202 The wind is added to the reports with the following convention: 1 to 2 knots 5 +/- 2 knots 10 +/- 2 knots 15 +/- 2 knots 20 +/- 2 knots 25 +/- 2 knots 30 +/- 2 knots 35 +/- 2 knots 40 +/- 2 knots 45 +/- 2 knots The direction is given by the orientation of the arrow and indicates where the wind is coming from. For the purpose of this thesis, only the wind, the cumulated amount of cloud N and the SST are important. The amount of cloud N is represented as follows: No cloud One okta or less 203 \ Vs u WW WWU Two oktas Three oktas Four oktas Five oktas Six oktas Seven oktas Completely overcast Sky completely obscured by fog, haze, smoke 204 Appendix B Features on Synoptic Weather Maps 1—j CO g> > « s u .s N O Si CD a. (1) "2 -y j= <« g T3 33 T3 - 01 01 «3 l-C I IH 0> > o J o> T3 <§ J= O . ^ £ efl O 0) C/5 > « o « IS o $ 6 .s s •6,1 3 * a? 0) 3 W J w CO o> 9> [ l £ to U £ 205 Appendix C Synoptic Weather Maps For Event 1 July 9 to August 8 (day 70 to day 100), 1990 206 207 208 209 210 2 1 1 2 1 2 2 1 3 2 1 4 215 217 2 1 8 2 1 9 220 2 2 1 2 2 2 2 2 3 224 2 2 5 2 2 6 2 2 7 2 2 8 2 2 9 230 2 3 1 2 3 2 2 3 3 2 3 4 2 3 5 2 3 6 2 3 7 Appendix D Synoptic Weather Maps For Event 2 July 3 to August 4 (day 64 to day 96),1992 2 3 8 Appendix D Synoptic Weather Maps For Event 2 July 3 to August 4 (day 64 to day 96),1992 2 3 8 2 3 9 240 2 4 2 2 4 3 2 4 4 2 4 5 2 4 6 2 4 7 2 4 8 2 4 9 250 251 2 5 2 2 5 3 2 5 4 2 5 5 2 5 6 2 5 7 2 5 8 2 5 9 260 2 6 1 2 6 2 2 6 3 2 6 4 2 6 5 2 6 6 2 6 7 2 6 8 2 6 9 270 2 7 1 Appendix E Synoptic Weather Maps For Event 3 June 29 to July 29 (day 60 to day 90), 1993 2 7 2 2 7 3 2 7 4 2 7 5 276 277 2 7 8 279 280 2 8 1 2 8 2 2 8 3 2 8 4 2 8 5 2 8 6 2 8 7 2 8 8 2 8 9 290 2 9 1 2 9 2 2 9 3 2 9 4 2 9 5 2 9 6 2 9 7 2 9 8 2 9 9 300 301 3 0 2 303 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099199/manifest

Comment

Related Items