Circulation and
Water Mass Formation in an Arctic Fjord
Elin Darelius
Geophysical Institute, University of Bergen UNIS, University Centre on Svalbard
August 15, 2003
This thesis is written in LaTex 2ε, reportstyle.
11 points text, double page format
This thesis is soon to be finished and if I haven’t done it before it is about time to thank all and everyone who has been involved. First of all, my supervisors Peter M. Haugan and Frank Nilsen who have been helpful and supportive all the way from Bergen to Billefjorden and back. I would also like to thank Pierre Jaccard for processing and helping me with ADCP data, Frank Cleveland for joining me in the quest for the missing water bottles, Ketil Eiane for bringing me on his expeditions and Dirk Notz for saving the RCM with his bare - red - hands.
Many thanks also to students and staff at UNIS who have helped me out in the field.
Thanks for spending long polar nights with CTD profiles and CTRL+F3 and for chilling your fingers while filling an endless number of water bottles. Only you know how cold that bottom water really is!
Finally I’d like to thank everyone who has brought a smile to my face when the hours and the articles were tedious and long - thanks for all the good laughters in the corridors and hallways of Geofysen and UNIS, in Odd, Drivhuset and studyroom 307 -without you I would never have made it!
I wrote finally before, but there is one more ”thank you” to add. It is a big one, and it should probably be in Swedish. So ”TACK” Karolina, for all discussions, comments, support and, a final finally, thanks for making me oatmeal!
Bergen, August 15, 2003
i
ii
1 Introduction 1
2 Billefjorden 3
2.1 Location . . . 3
2.2 Climate . . . 4
2.2.1 Freshwater input . . . 6
2.2.2 Ice cover . . . 7
2.3 Previous and ongoing work . . . 8
3 Instruments and Methods 9 3.1 CTD . . . 9
3.1.1 Calibration . . . 10
3.1.2 Ice-cores . . . 10
3.2 ADCP . . . 10
3.2.1 ADCP zones . . . 11
3.2.2 ADCP Data Processing . . . 11
3.3 RCM . . . 12
3.4 Fieldwork and cruises . . . 13
4 Field data 15 4.1 CTD . . . 15
4.2 ADCP . . . 15
4.3 RCM . . . 21
4.4 Weather data . . . 22
4.5 Ice Observations . . . 22
4.5.1 Ice distribution . . . 22
4.5.2 Ice-cores . . . 24
4.6 Tidal predictions . . . 25
5 Hydrography and seasonal changes 26 5.1 Theory . . . 26
5.1.1 Freshwater content . . . 26
5.1.2 Heat Content . . . 26
5.2 Discussion . . . 27 iii
iv CONTENTS
6 Fjord Circulation 32
6.1 Theory . . . 32
6.1.1 Estuarine circulation . . . 32
6.1.2 Rotational Effects . . . 33
6.1.3 Wind effects . . . 34
6.1.4 Up-welling at glacier front . . . 36
6.2 Discussion . . . 36
7 Sill Flow 40 7.1 Theory . . . 40
7.1.1 Hydraulic control . . . 40
7.1.2 Mixing . . . 42
7.2 Discussion . . . 42
8 Tides in a fjord 45 8.1 Theory . . . 45
8.1.1 Internal tides . . . 46
8.1.2 Tidal choking and Jets . . . 46
8.1.3 Tides under ice . . . 46
8.1.4 Tidal constituents and harmonic analysis . . . 47
8.2 Discussion . . . 47
9 Ice formation and convection 51 9.1 Theory . . . 51
9.1.1 Freezing . . . 51
9.1.2 Salt and Brine . . . 53
9.1.3 Convection . . . 53
9.2 Discussion . . . 54
9.3 One-dimensional ice model . . . 56
9.3.1 Model results . . . 57
9.3.2 Discussion . . . 57
10 Deep water renewal 61 10.1 Theory . . . 61
10.1.1 Deep water renewal . . . 61
10.1.2 Vertical Diffusion . . . 61
10.2 Discussion . . . 63
11 Summary and Conclusion 65 11.1 Future work . . . 66
A Equations and Parametrization 67 A.1 Drag coefficient . . . 67
A.2 Salt and Heat transport . . . 67
A.3 Sensible Heat flux . . . 67 A.4 Sea ice properties . . . 68
B Instrumentation 69
B.1 CTD -SBE 911plus . . . 69 B.2 ADCP . . . 70 B.3 RCM7 . . . 70
C Calibration of CTD data 72
C.1 Jan Mayen -December 2002 . . . 72 C.2 H˚akon Mosby -October 2002 . . . 72
D Additional Data 76
D.1 CTD . . . 76 D.2 ADCP data . . . 76 D.3 RCM7 . . . 76
E Harmonic analysis 81
E.1 Harmonic analysis in Billefjorden . . . 81
F One dimensional Ice/Convection model 83
G Estimation of Vertical Diffusivity 85
Bibliography 88
Chapter 1
Introduction
It is generally accepted that the Arctic is an important factor in the world’s climate. Mean- while, the Arctic itself is very sensitive to variations in temperature, precipitation and wind- patterns. The feedback-loops are numerous and the processes interacting complicated and poorly understood. In order to better monitor and predict changes within the Arctic Ocean it is essential to improve the understanding of the system as a whole, and so of the Arctic fjords. A fjord is not a locked system; it influences, and is influenced by the ocean outside.
The interest for fjords on Svalbard was first woken by Helland-Hansen and Nansen, who in a pioneer study discussed the nature and origin of the water in the fjords (Helland-Hansen
& Nansen (1912) and Nansen (1915)). Recently, much attention has been focused on Stor- fjorden east of Svalbard, where formation and outflow of dense bottom water have been observed by e.g. Quadfasel, Rudels & Kurz (1988). UNIS (University Courses on Svalbard) have centered their work on Van Mijenfjorden, resulting in a number of Master’s theses and student reports. As an example, Kangas (2002) studied seasonal variations in temperature and salinity while Skardhamar (1998) dealt with circulation in the fjord. The circulation in Kongsfjorden-Krossfjorden was studied by Ingvaldsen et al (2001) while Svendsen et al (2002) describes the physical environment of the same fjord system.
The objective of this Master’s thesis is to give a physical description of Billefjorden and to explore the mechanisms controlling circulation and water mass formation within an arctic fjord. Hydrographical and hydrodynamical data have been collected throughout the autumn and winter, and the results are presented and discussed in relation to traditional fjord theories.
The thesis is so to say an ”Introductory course” in the oceanography of Billefjorden and will hopefully be helpful when planning and conducting further marine studies in the area.
The thesis is organized in eleven chapters, starting out with a general description of the area and its climate in Chapter 2 and continuing with a description of field work and instrumentation in Chapter 3. Field data and other data-sets are presented in Chapter 4.
1
The thesis covers five major topics:
• Hydrography and seasonal changes
• Fjord circulation
• Sill flow
• Tides in a fjord
• Ice formation and convection
• Deep water renewal
which are treated separately in Chapter 5-10. These chapters all consist of two parts, a preliminary theory part which is followed by a discussion of the data (from Chapter 4) in relation to the theories presented. Further calculations and analyses, suggested by the theories, are performed within the discussion or in the appendix. A one-dimensional model simulating ice-growth and convection is included in Chapter 9. Chapter 11 comprises a summary and suggestions to future work.
Chapter 2
Billefjorden
2.1 Location
Spitsbergen is the largest island in the Arctic archipelago of Svalbard. The archipelago is situated between 76◦ and 81◦ North and 10◦and 34◦ East, bordering the Arctic Ocean in the north, the shallow Barents Sea in the east and the Fram Strait in the west. The northernmost branch of the Gulf stream, called the West Spitsbergen Current (WSC), follows the shelf- break just west of Spitsbergen. The warm and saline water flowing northward keeps the west
6oE
12oE 18oE 24oE 30oE
75o N 76o
N 77o
N 78o
N 79o
N 80o
N 81o
N 82o
N
Arctic Ocean
Barents Sea Fram Strait
Stor Fjorden Billefjorden
Isfjorden Longyearbyen
.
Sassen−Tempel fjordenGreenland Sea Kongsfjorden
Hornsund
Figure 2.1: Svalbard 3
coast more or less ice free throughout the winter. The west coast of Spitsbergen is indented with numerous fjords, the largest of them being Isfjorden. Billefjorden1 is the innermost side-fjord of Isfjorden, see figure 2.1.
The Norwegian Polar Institute made detailed bathymetric investigations of Billefjorden in 1974 and a simplified version of their map (Norwegian Polar Institute 1974) can be found in figure 2.2.
The fjord is about 30 km long and 5 to 8 km broad. It is a two-silled fjord, with an outer sill-depth of about 70 meter and an inner sill-depth of 50 meter.
In between the sills, to the west, there is a deep region with a maximum depth of 230 m. The eastern side of the outer basin is shallow with depths between 15 and 30 meters. The outer basin is about 7 km long and 8 km wide.
The inner basin, on which the work in this thesis is focused, extends some 17 km before it divides in two short ”arms”. The western arm, Petuniabukta, is relatively shallow (40 to 60m) while the eastern arm, Adolfbukta, is deeper with a maximum depth of 215 m.
The inner basin has steep topography and is approximately 160 m deep with a shallower area (95 m) in the middle.
Billefjorden is surrounded by rather steep mountains, of whichPyramiden on the western coast is probably the most characteristic. The abandoned Russian settlement carrying its name is found just below the mountain, in Mimerbukta, while the grey huts of Brucebyen are located across the fjord. Both settlements are remnants of earlier mining projects. The Nordenskjøld glacier termins in Adolfbukta.
2.2 Climate
The climate on Svalbard is relatively mild, considering its northerly position. Ocean cur- rents and general wind patterns help out to bring the mean annual temperature up to−6◦C (Hagen, Liestøl, Roland & Jørgensen 1993). The low pressure area near Iceland generally produces southerly winds transporting warm air from lower latitudes in over Svalbard. North of the archipelago the high pressure area over Greenland and the Arctic Sea causes northerly winds that bring down cold arctic air. Extremely large fluctuations in weather and temper- ature are observed on Svalbard and they are related to motions of the polar front2. This is most noticeable in winter, when the temperature difference between the two air masses is largest (Førland, Hanssen-Bauer & Nordli 1997).
DNMI (the Norwegian Meteorological Institute) has several weather stations located in the archipelago and climatic data from these are presented in Førland et al. (1997). The mean temperature in January-March is−15◦ C. July is the warmest month, with mean tem- peratures of 5−6◦C (Hagen et al. 1993).
1Billefjorden was named after the Dutch whaler Cornelius Claeszoon Bille who was active in the area around 1675. The name was originally used on Adventfjorden, but Nordenskjøld and Dun´er transferred it to its present place (Orvin 1991).
2The boundary between the cold arctic air and warm, southern air
2.2. CLIMATE 5
Figure 2.2: Bathymetry and place names in Billefjorden, after Norwegian Polar Institute (1974).
The station at Svalbard airport, situated in Longyearbyen, is the weather station situated closest to Billefjorden and mean temperature data (1961-1990) from this station is presented in figure 2.3. The mean annual temperature at Svalbard airport is -6.7◦C (Førland et al. 1997).
A) B)
J F M A M J J A S O N D
−20
−15
−10
−5 0 5 10
Temperature, [° C]
J F M A M J J A S O N D
0 5 10 15 20 25 30
Precipitation, [mm/month]
Figure 2.3: A) Mean temperature and B) precipitation at Svalbard Airport (Førland et al, 1997) There is normally little precipitation on Svalbard. Mean annual values from the DNMI weather stations are in the range of 190-525 mm. The lowest annual value (190 mm) is from the weather station at Svalbard airport (Førland et al. 1997). Mean precipitation on the western coast of Svalbard is 400mm (Hagen et al. 1993).
Topography greatly influences the precipitation rate, and precipitation has been shown to increase with height (Førland et al. 1997).
Data from the DNMI weather stations in Spitsbergen show that the prevailing wind direction is from land towards sea, along the fjord or valley where the station is situated (Førland et al. 1997).
2.2.1 Freshwater input
Fjords on Spitsbergen receive freshwater from four sources:
• Tidal glacier ablation and calving
• Direct precipitation on the fjord surface
• Melting of fast ice entering the fjord
• Land/riverine outflow
2.2. CLIMATE 7 There are no investigations concerning the freshwater input to Billefjorden, but the sit- uation is probably similar to other fjords on Svalbard where such investigations have been performed.
In Hornsund, the southernmost fjord on Spitsbergen (see figure 2.1), glacial ablation contributes with 86% of the freshwater input while only 8% comes from rivers (Weslawski, Jankowski, Kwasniewski, Swerpel & Ryg 1995), see figure 2.4. In June, the melt of advected pack-ice is here just as important as glacier ablation. More than 95% of the freshwater enters Hornsund between June and September.
Svendsen et al. (2002) estimated the freshwater input to the Kongsfjorden - Krossfjorden system (see figure 2.1 and found that more than 90 % of the freshwater entered the fjord system during the three summer months.
0 100 200 300 400 500 600
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Freshwater input x 106 m3 /month
Total Glacier River Snow/Ice
Figure 2.4: Freshwater contribution in Hornsund, from Weslawski et al (1995).
With a drainage area of 907 km2and an annual precipitation of 400 mm (Hagen et al. 1993) the total fresh water input to Billefjorden can be estimated (assuming that the 48 glaciers in the area are in mass balance) to about 360∗106 m3 per year. Most of the freshwater probably enters the fjord during the summer months, but some discharge from the Nordenskjøld glacier can be expected also during winter and spring (Jan Ove Hagen, personal communication).
2.2.2 Ice cover
The ice cover in Billefjorden is seasonal, i.e. all sea-ice in the fjord melts during summer.
The ice cover normally consists of fast ice which is formed locally, but drifting ice may enter from Isfjorden. Fresh ice from the calving Nordenskjøld glacier that is trapped in the fjord
at freeze-up, freezes into the sea-ice.
The ice cover is normally weak or nonexisting in the sill area.
In 2001 freezing did not start until February while the fjord was covered with ice from January to the end of June in 2002 (Arnkvern 2002).
2.3 Previous and ongoing work
Billefjorden has not been given much attention in ”oceanographical” literature and little ob- servational data is available.
Helland-Hansen & Nansen (1912) mentions stations occupied by ”the Swedish oceanogra- phers” (de Geer) in ”Klaas Billen Bay” as early as 1908.
The University of Tromsø visited the fjord in the period 1979-1985 (August/September) as a part of their yearly cruises in the Svalbard area. The main interest of these cruises were marine biology, but hydrographical data was also collected and is presented in Normann &
Pettersen (1984) and Normann (1986).
Tutorial cruises with geophysics students from UNIS (the University Centre on Svalbard) have collected data in the area and the results are presented in student rapports, e.g. Nilsen (2002).
Hydrological data was also collected by the University of Bergen (UiB) in cooperation with UNIS during September and December 1998 (unpublished). Data from these cruises is pre- sented in the appendix, section D.1.
At present, UNIS (Ketil Eiane) is running a biology project in Billefjorden. As a part of the project hydrological data was collected during the spring of 2002 and 2003.
Chapter 3
Instruments and Methods
The data used in this thesis were collected during five cruises in the fjord with R/V H˚akon Mosby and R/V Jan Mayen, starting on 17 August, 22 September, 11/18 October and 11 De- cember 2002 respectively. Data was also collected from the ice on during the spring, 15/24/26 March and 11/25 April 2003.
Hydrographical data was collected using a CTD (Conductivity-temperature-depth-profiler) and current measurements was performed with an ADCP (Acoustic Doppler Current Pro- filer) and a RCM (Recording Current Meter). The instruments are presented in the sections below and in the appendix.
Weather observations from DNMI and tidal predictions from Longyearbyen (Sjøkartverket (2001) and Sjøkartverket (2002)) are available as well as satellite images fromwww.sarepta.org (2003) showing ice distribution during the winter of 2003.
3.1 CTD
Autumn 2002
During the cruises with R/V H˚akon Mosby and R/V Jan Mayen, a total of 320 CTD profiles were taken, using a Seabird Electronic Inc. SBE911plus sonde. The CTD is equipped with sensors for conductivity, temperature and pressure and returns salinity and temperature pro- files when lowered through the water column, see section B.1 in the appendix for a description of the instrument and the sensors. Occasionally a transmissiometer and/or an oxygen sensor was also attached to the sonde, see table 3.1. The sonde was mounted on a SBE32 Carousel, containing a varying number of 10 liter Niskin bottles for collection of water samples. The position and time of occupation of the stations were determined from the on-board differen- tial GPS system, Seapath 200, with corrections from FUGRO Seastar 3000.
Water samples for calibration were collected (except for on the August cruise) following the standard Unesco (1988) procedures.
The location of the CTD sections are shown in figure 3.1 A-B and their time of occupation 9
in table 3.1. The number of stations in each section and their exact location vary somewhat from time to time.
Spring 2003
During the field work in March 2003 five CTD stations were occupied from the fast sea ice using a Seabird Electronics Inc. 19 SEACAT Profiler. The SEACAT resembles the SBE911 plus described in the appendix, section B.1, but it is ”free-flushed” i.e. it has no pump controlling the time lag between conductivity and temperature measurements, see Tverberg (1997) for further information on the instrument. The stations are located as indicated in figure 3.2.
When occupying a station from the ice the following procedure was followed:
Upon arriving to the station, a hole was drilled in the ice and the depth was measured.
Meanwhile, the CTD was heated1in order to prevent ice from forming on the cold instrument when lowering it into the water. The hole was carefully cleared from slush and ice with a sieve before the instrument was taken from the heat and rapidly lowered down to ten meters depth. The instrument was kept at ten meters depth for approximately one minute so that eventual ice on the sensors would melt, before it was raised to the surface and the actual profiling started. The instrument was lowered and mounted manually with a winch, at an approximate speed of 1 m/s and data was internally recorded. Water-samples were collected using a Nansen bottle.
3.1.1 Calibration
Conductivity sensors are known to drift and calibration, i.e. comparison with collected water samples, are to be performed for each cruise. The data sets have been calibrated (see Nilsen (2002) and section C in the appendix) with the exception of the cruise with Jan Mayen in August 2002 and the data collected during the spring of 2003.
3.1.2 Ice-cores
Three ice-cores were taken in relation to CTD-stations (1,3 and 8, see figure 3.2 A) in March 2003. A temperature profile of the ice was obtained by drilling holes every fifth centimeter and inserting a thermometer. Air and snow temperature was also recorded. The core was cut in pieces of about 10 cm which were brought to Longyearbyen for analysis. The salinity was determined using a WTW Multi 340 and a TetraCon 325 standard conductivity cell.
3.2 ADCP
The ADCP is an acoustic current meter that provides a vertical current profile using Doppler principles, see section B.2 in the appendix.
1using a generator and a heater
3.2. ADCP 11
Date Sensors 1 2 3 4 5 Sill A-E Stations
Mosby #1 17 Aug, 2002 C,T,P x x
Mosby #2 22 Sep C,T,P,O x x x x x
25 Sep C,T,P,O x
26 Sep C,T,P,O x x x x x
1 Oct C,T,P,O x
Mosby #3 11-12 Oct C,T,P,O x
16 Oct C,T,P,Tr x x x x
Mosby #4 20 Oct C,T,P,O,Tr x
Jan Mayen 11 Dec C,T,P x
24/26 Mar, 2003 C,T,P 1*,3*,5,8*,14
Biologists 15 Mar, 11/25 Apr C,T,P 1
Table 3.1: CTD sections taken in Billefjord; C=Conductivity, T=Temperature, P=Pressure, O=Oxygen, Tr=Transmissivity, F=Flourescence. *ice-core
During the cruise with R/V H˚akon Mosby on 18-20 October 2002, the Vessel Mounted Narrow Band 150kHz ADCP (manufactured by RD Instruments (RDI)) was used to obtain velocity profiles across the fjord, see figure 3.2 A. The ship was driving back and forth at
∼4 knots, giving approximately two transects per hour. Starting on 19 October (20:00), the procedure changed slightly as a CTD station (see figure 3.2 A) was included. The ADCP was from then on running in only one direction and the CTD-station was taken on the way back, resulting in only one transect per hour. The ADCP measurement was ended at 20th of October 2002, after a total of 61 ”runs” or transects across the fjord.
A cell depth of 8 meters and a blanking distance2 of 12 meters was used, giving that the first measurement are centered at 16 meters depth, the next at 24 meter and so on.
3.2.1 ADCP zones
To look at spatial variations, the ADCP-transect was divided in four zones. The partition was performed along the longitudes, with zone one corresponding to 16◦300−16◦330, zone two to 16◦330−16◦360 and so on. The zones are shown in figure 3.2 B together with the bottom topography along the section.
3.2.2 ADCP Data Processing
In accordance with Emery & Thomson (1998) data from the lowermost 20 percent of the depth was disregarded, since those measurements are affected by bottom topography.
The ADCP measures the velocity relative to the ship, and the absolute velocity was obtained by adding the ship-velocity. The ship velocity can be determined either from GPS
2To avoid non-linear effects the area closest to the transducers is disregarded
A) B)
45’ 16oE 15’ 30’ 45’ 17oE 24’
28’
78oN 32.00’
36’
40’
44’
1.
2.
3.
4.
5.
16oE 10’ 20’ 30’
26’
28’
78oN 30.00’
32’
34’
36’
A
C B
D E
Figure 3.1: CTD sections in Billefjord: A) Section 1-5 and B) the ”sill-sections”.
positioning (The ADCP was connected to a Seatex Seapath 200 DGPS, which was corrected by FUGRO Seastar 3000) or through bottom tracking, but since GPS-positioning in the Svalbard area is poor (Pierre Jaccard, personal communication) the latter was preferred.
The data was initially averaged over one minute, with 35 to 45 pings in each ensemble.
According to Emery & Thomson (1998) this set up would give a standard deviation of 3 cm/s for a stationary instrument, while it would be approximately 6 cm/s for a vessel mounted system using bottom tracking (Pierre Jaccard, personal communication). When used in calculations, several measurements have been averaged together. For a ”run” across the fjord consisting of ∼30 measurements the standard deviation is ∼1 cm/s. (For a run across a section the deviation is ∼2 cm/s)
3.3 RCM
A rotor RCM7 (Recording Current Meter), manufactured by AAnderaa Instruments A/S, provides a current record from the position and depth were it is placed. The current mea- surements are mechanical, and due to friction currents speeds lower than 2 cm/s can not be measured accurately, (www.esica.com/products/aanderaa/rcm7.htm 2003). The instrument is presented in detail in the appendix, section B.3.
A mooring consisting of two RCM7s was deployed from the fast ice of Billefjorden on 12 March 2003 and recovered 15 days later, on 28 March. The position of the mooring is shown in figure 3.2 A. The current meters were placed at 10 and 40 meters depth respectively and they were set to a recording interval of ten minutes.
3.4. FIELDWORK AND CRUISES 13
A) B)
45’ 16oE 15’ 30’ 45’ 17oE 24’
28’
78oN 32.00’
36’
40’
44’
CTD 8/RCM
1 3 5
14
ADCP
0 0.5 1 1.5 2 2.5 3
0
50
100
150
Depth, [m]
[km]
Depthprofile, ADCP transect 2
Zone 1 Zone 2 Zone 3 Zone 4
Figure 3.2: A) Location of the ADCP transect, the RCM mooring and spring/ADCP CTD stations and B) zones and bottom topography along the ADCP transect.
3.4 Fieldwork and cruises
• 17 August, 2002
Biology students from UNIS visited the area with R/V H˚akon Mosby. CTD section 1 and 5 was occupied. Unfortunately, no water samples were collected and the data has not been calibrated. However, un-calibrated bottom salinities from this cruise agree well with salinities from later, calibrated, cruises.
• 22 September-1 October, 2002
Geophysics students from UNIS visited Isfjorden and Billefjorden several times during a cruise with R/V H˚akon Mosby. CTD sections 1-5 were occupied repeatedly and water bottles for calibration were collected. Calibration was performed by the students at UNIS (Nilsen 2002).
• 11-16 October, 2002
During this cruise with R/V H˚akon Mosby the sill sections, see figure 3.1 B, were occupied twice. The along fjord section (A) was occupied during ebb and flood, while the transverse sections (B-E) were occupied during maximal tidal flow. Three stations at the outermost transverse section (E) had to be left out due to shortage of time.
On returning to Billefjorden after a session in Storfjorden CTD section 1,2,3 and 5 was occupied. Water samples for calibration and chemical analysis were collected and
analysed by Fredrik Svendsen, UiB.
• 18-20 October, 2002
During 36 hours a total of 61 ADCP transects were run across the fjord, just inside the innermost sill, see figure 3.2. Towards the end a CTD station was included in the procedure and it was occupied 8 times.
Before leaving Billefjorden a final CTD-section, section 1, was occupied.
Water bottles for calibration were collected analysed by Fredrik Svendsen, UiB.
• 11 December 2002
R/V Jan Mayen visited Billefjorden and CTD-sections 1 and 5 were occupied. Water samples were collected at each station and they were analysed at UNIS by Elin Darelius.
No sea-ice was observed.
• 12 March, 2003
A mooring with two RCM7s, (10 and 40 m depth) were deployed from the ice. The fjord was accessed by scooter.
• 24 and 26 March, 2003
A total of five CTD stations in the inner, ice-covered part of the fjord were occupied.
Three ice-cores and water samples for calibration were collected. The fjord was accessed by helicopter.
• 28 March, 2003
The RCM mooring was recovered.
• 15 and 28 March, 11 and 25 April, 2003
Biologist students from UNIS visited the fjord and a CTD profile was obtained from the inner part of the fjord (station 1).
Chapter 4
Field data
Data acquired in field is presented and briefly commented in this chapter while further cal- culations and derived results are found in the discussion parts of the following chapters.
4.1 CTD
Along-fjord (section 1 in figure 3.1 A) salinity and temperature profiles from August, October and December are presented in figure 4.1 and 4.2, while across-fjord profiles (section 2 in figure 3.1 A) from September are presented in figure 4.3.
The bottom topography is interpolated from the depth at the included stations and is not realistic. The sills have been added at their approximate locations and they are indicated with arrows.
Data from the spring of 2003 are shown in figure 4.4. The salinity data obtained from the UNIS biology project1 is not calibrated and the values are obviously too high.
Along-fjord sections are presented with the inner part of the fjord to the right, and across- fjord sections with western shore to the left (i.e. looking up-fjord).
4.2 ADCP
A total of 61 ADCP transects were made across the fjord 18-20 October 2002 (see figure 3.2 for location).
Figure 4.5 A shows the mean (two tidal periods) along fjord velocity and reveals a two- layered circulation with outflow in the surface (above 32 m) and inflow at depth (32-72m).
The inflow is on the order of 10 cm/s while the measured outflow is weaker. The mean flow is oriented at 30◦ angle to the fjord.
The CTD stations occupied during the current measurements show the same layered structure as the ADCP data, see figure 4.5 B. The out-flowing surface water is well-mixed and extends down to just below thirty meters depth. It is fresher and colder (S'33.3 psu,
1Collected with a STD
15
A)
B)
29.5 30.5 31.5 32.5 33.5 34.5
50
100
150
200
C)
29.5 30.5 31.5 32.5 33.5 34.5
0
50
100
150
29.5 30.5 31.5 32.5 33.5 34.5
50
100
150
Figure 4.1: Salinity [PSU] sections from Billefjorden (section 1), A) 17 August, B) 20 October and C) 11 December 2002. Arrows indicate locations and depths of sills.
T'2◦C) than the inflowing layer below (S'33.5-34.5 psu, T'3◦C). Below the inflowing layer which extends down to 65 m, cold and saline bottom (S'34.6 psu, T'-1.9◦C) is found.
4.2. ADCP 17 A)
B)
−2 −1 0 1 2 3 4 5
50
100
150
200
C)
−2 −1 0 1 2 3 4 5
0
50
100
150
−2 −1 0 1 2 3 4 5
50
100
150
Figure 4.2: Temperature [◦C] sections from Billefjorden (section 1), A) 17 August B) 20 October and C) 11 December 2002. Arrows indicate locations and depths of sills.
When looking at the mean from each run across the fjord and subtracting a twelve hour
A) B)
29.5 30.5 31.5 32.5 33.5 34.5 20
40 60 80 100 120 140
−2 −1 0 1 2 3 4 5
20 40 60 80 100 120 140
C) D)
29.5 30.5 31.5 32.5 33.5 34.5 20
40 60 80 100 120 140
−2 −1 0 1 2 3 4 5
20 40 60 80 100 120 140
Figure 4.3: Transverse salt and temperature distribution across the fjord (section 2) A-B) 22 Septem- ber and C-D) 26 September 2002. 2002.
running mean a sinusoidal signal with an amplitude of 2 cm/s and a period of about 12 hours is apparent above sill depth. The signal is present in both U (East-West) and V (North- South), see figure 4.6 A-B. The signal is strongest in zone 3 and 4 (see figure 3.2 B for location), where the amplitude is about 6 cm/s, while it is weaker or nonexisting in zone 1 and 2, see figure D.3 in the appendix.
Removing this signal from the measurements by looking at a twelve hour running mean (see figure D.4 in the appendix) it is obvious that the back ground current is not stationary but evolving on a longer time scale. As an example, the current at 40 meters depth is turning from across fjord to along fjord in zone 2-4. The inflow is increasing in magnitude while the outflow remains more or less constant.
4.2. ADCP 19
A) B)
−2 0 2
0
50
100
150
200
Temperature [oC]
Pressure [dbar]
12/11 03/15 03/26 04/11 04/25
32 34 36
0
50
100
150
200
Salinity [PSU]
pressure [dbar]
12/11 03/15 03/26 04/11 04/25
−2 −1.5 −1
0
50
100
150
200
Temperature [oC]
Pressure [dbar]
Sta 1 Sta 3 Sta 5 Sta 8 Sta 14
34 34.5 35
0
50
100
150
200
Salinity [PSU]
pressure [dbar]
Sta 1 Sta 3 Sta 5 Sta 8 Sta 14
Figure 4.4: Salt and temperature profiles from A) the inner part of the fjord (station 1) from December 2002 to April 2003 and B) from stations 24-26 Mars, 2003
A) B)
−5 0 5 10
16 24 32 40 48 56 64 72 80 88 96 104
Velocity [cm/s]
Depth [m]
−2 0 2
0 20 40 60 80 100 120
Temperature °C
Pressure (dbar)
33.5 34 34.5 Salinity (psu)
27 28
0 20 40 60 80 100 120
Sigma theta
Figure 4.5: A) Mean (two tidal periods) along fjord velocities from ADCP measurements 18-20 October, positive values indicating flow into the fjord and B) salinity, temperature and density profiles from the CTD station occupied in relation to the ADCP transects, October 20, 2002.
A) B)
00:00 06:00 12:00 18:00
−4
−3
−2
−1 0 1 2 3 4
Velocity, [cm/s]
16 m 24 m 32 m 40 m
00:00 06:00 12:00 18:00
−4
−3
−2
−1 0 1 2 3 4
Velocity, [cm/s]
16 m 24 m 32 m 40 m
Figure 4.6: Deviations from a twelve hour running mean in A) U and B) V.
A)
03/14 03/16 03/18 03/20 03/22 03/24 03/26 03/28
5 cm/s
B)
03/14 03/16 03/18 03/20 03/22 03/24 03/26 03/28
5 cm/s
Figure 4.7: Twelve hour running mean velocity from RCM7 at A) 10 m and B) 40 m
4.3. RCM 21
4.3 RCM
The RCM7-mooring provides velocity and temperature data from 10 and 40 meters depth in the period 12-28 March 2003. (The salinity measurements have been disregarded as the conductivity sensors were not to be trusted.)
Velocities are low, with 31(40) % of the data at 10(40) m being below the threshold value for the instruments (2 cm/s), see figure D.5 in the appendix.
The twelve hours running mean are shown in figure 4.7 A-B. Velocities are on the order of 5 cm/s and the mean transport is to the southwest.
When subtracting a twelve hour running mean a sinusoidal signal with an amplitude of 1-2 cm/s and a period of about 12 hours can be surmised in the V-component of both records (10 and 40 meters), see figure 4.8 A-B. No such signal is apparent in the U component.
A)
B) 03/15 03/17 03/19 03/21 03/23 03/25
−4
−2 0 2 4
Velocity, [cm/s]
10 m 40 m
03/15 03/17 03/19 03/21 03/23 03/25
−4
−2 0 2 4
Velocity, [cm/s]
10 m 40 m
Figure 4.8: Deviations from a twelve hour running mean in A) the U component and B) the V component, from RCM7
The temperature at 10 m is at or close to the freezing point while the temperature at 40 meters shows bursts of ”warmer” water on a few occasions, see figure 4.9. The warmest burst occurs in the end of the measured period (27-28 March, 2003) when the temperature reaches
∼ −1.3◦C for two days. The CTD-profiles from the inner part of the fjord from 26 March 2003 show a layer with the same temperature at around 40 meters depth, see figure 4.4.
03/09 03/16 03/23 03/30
−1.8
−1.7
−1.6
−1.5
−1.4
−1.3
Temperature, [o C]
10 m 40 m
Figure 4.9: Temperature at 10 and 40 m depth, from RCM7 March 2003
4.4 Weather data
Meteorological data from the Billefjorden area is available only from cruises with H˚akon Mosby. Data from the weather station at Svalbard airport (∼ 30 km from Billefjorden) has been obtained from DNMI. Wind observations from two of the cruises with R/V H˚akon Mosby to Billefjorden are presented in figure 4.10 A-B and temperature observations from Svalbard airport in figure 4.11.
4.5 Ice Observations
4.5.1 Ice distribution
Satellite images over the area were downloaded from www.sarepta.org (2003) and examples are presented in figure 4.12.
The satellite images show signs of ice already the 2-3 January (2003) and by 20 January the whole fjord is covered with ice. On the 5 February, the ice cover starts to retreat and it reaches a minimum extent around the 22 February. At this point the ice does not extend far beyond Pyramiden and Mimerbukta.
The satellite image from 28 February shows that ice is again forming and on 11 March the fjord is totally covered with ice.
On arriving to the fjord on 12 March the fjord ice was flat and no ridges were observed. A crack in the ice, about 30 cm wide and covered with 20 cm of new ice, was observed at the approximate location of the minimum ice extent. Inside of the crack, which extended from one side of the fjord to the other, the ice was about 80 cm thick while it measured only 30 cm on the outside. The snow cover differed from one side of the crack to the other, inside of the crack, there were 10 cm of snow while only 5 cm was observed outside of the crack.
The ice seems to disappear from the outer basin around 23 March while the inner basin remains covered at least to 25 April, when the last satellite image was down-loaded.
4.5. ICE OBSERVATIONS 23
A) B)
5 10 North 15
South
West East
22 sep −02 25 sep −02
5 10
15 North 20
South
West East
18−20 oct
Figure 4.10: Wind observations during cruises with H˚akon Mosby, A) 22-25 September and B) 18-20 October 2002.
Aug Sep Oct Nov Dec Jan Feb Mar Apr May
−30
−25
−20
−15
−10
−5 0 5 10 15
2002−2003 Temperature, [ ° C]
Figure 4.11: Temperature observations at Svalbard Airport 2002-2003, from DNMI.
Figure 4.12: Satellite images of the Billefjorden area, from www.sarepta.org (2003)
4.5.2 Ice-cores
Temperature and salinity profiles from the ice cores taken in Billefjorden in March 2003 are presented in figure 4.13 A-B. The salinity-profiles show a c-shape, with higher salinities at the surface and bottom of the ice-core. The bulk salinity is about 7 psu.
A) B)
−20 −15 −10 −5 0
−80
−70
−60
−50
−40
−30
−20
−10 0 10 20
Temperature, [oC]
Depth, [cm]
ICE SNOW AIR
Sta 8 Sta 3 Sta 1 Air Temp
5 6 7 8 9
−80
−70
−60
−50
−40
−30
−20
−10 0 10 20
Salinity, [PSU]
Depth, [cm]
Sta 8 Sta 3 Sta 1
Figure 4.13: A) Temperature and B) salinity profiles from ice cores, 24-26 March 2003
4.6. TIDAL PREDICTIONS 25
4.6 Tidal predictions
Tidal predictions are not available for Billefjorden but predictions for Longyearbyen during the period of interest was found in Sjøkartverket (2001) and Sjøkartverket (2002).
Fitting the predicted water levels to a curve on the form
F(t) =Ki+Ai∗cos(ωi(t−φi)) (4.1) where each half period is numbered i and then differentiating, gives an indication of the tidal current phase in Longyearbyen. The current magnitude depends on local parameters and can not easily be estimated. When presented together with current measurements for comparison, the magnitude of the ”Longyearbyen tide” has been scaled. The magnitude is chosen such that the current would produce a change in water level in the fjord similar to the one observed in Longyearbyen.
Hydrography and seasonal changes
5.1 Theory
In literature, (e.g. Gade (1986) and Farmer & Freeland (1983)) the fjord water is often divided in three layers; an upper brackish layer, an intermediate layer and finally the deep basin water which is shielded from the sea outside by the sill, see figure 6.1.
Temperature, salinity and thickness of the layers change throughout the year due to atmo- spheric forcing, variations in fresh water input, ice freezing and advection.
5.1.1 Freshwater content
The freshwater content can be defined as the amount of freshwater needed to obtain the observed salinity profile when mixing with water of a certain reference salinity, Sref. The freshwater content can easily be found from a CTD-profile using equation 5.1.
Qf =XSref −Sz
Sref ∆Z (5.1)
5.1.2 Heat Content
The heat content can be defined in two ways -either by referring to a constant reference temperature or by referring to the salinity and pressure dependent freezing. The latter gives the available heat, i.e. the heat that can be removed from a volume of water before it starts to freeze. The available heat can be found from equation 5.2.
Havailable=X
ρ∗Cp∗(T(z)−Tf reeze(z))∗∆z (5.2) The heat content of the fjord diminishes as water is cooled during the autumn and winter.
Gade, Lake, Lewis & Walker (1974) found that half of the total annual heat loss in Cambridge bay (Canadian Arctic archipelago) occurred during the month before freeze-up.
Freezing can start when the available heat is removed from the surface layer, even if heat is contained at depth.
26
5.2. DISCUSSION 27
5.2 Discussion
Seasonal changes
The CTD-data presented in figure 4.1-4.4 show how the hydrography of Billefjorden changes throughout the autumn, winter and spring. The three-layered model described by e.g. Gade (1986) and Farmer & Freeland (1983) is well applicable in the autumn when the brackish, intermediate and bottom layer is well defined, but not in the spring when ice freezing and convection have made the water column nearly homogenous.
The observations are described below.
• 17 August (2002)
In August the fjord is strongly stratified with a fresh and warm (S=30.5-31.5 psu, T=5◦C) surface layer extending down to about 25 meters depth. Below sill level (50 m) the water is saline and at or close to its freezing temperature (S'34.7psu, T'-1.9◦C).
The outer, deep basin is filled with just as cold but slightly fresher water. Here, the cold water is found below 125 meters depth, or well below the outer sill-level (70 m).
• 22-26 September
In September, the fjord has started to cool off from above. Surface temperatures are between 2.5-3.5◦C, while temperatures around 30 meters are higher (T'4-5◦C). The salinity of the surface water has increased to ∼32.5 psu.
The properties of the inner basin water are unchanged. However, in the outer basin the cold water is now to be found only below 150 meters depth.
From the transverse sections (see figure 4.3) it is obvious that the surface-water is displaced to the east (right) on 22 September and to the west (left) on the 26 September.
This will be discussed further in the discussion of wind effects in section 6.2.
• 20 October
In October, the cooling has reached down to 50 meter and the well mixed surface layer has a temperature of∼2◦C and a salinity of ∼33.3 psu. At the innermost station the salinity is higher, S∼33.5 psu.
In the outer basin, there is a strong pycnocline at 30 meters depth, separating the surface water from the warmer water at mid-depth, where temperatures are between 3 and 4◦C. This water is noticeably warmer than it was in September, when temper- atures at this depth were 2-3◦C. A strong thermocline at 125 meters depth separates the warm intermediate water from the cold bottom water.
The basin water of the inner basin is unchanged at depth, while warm water has spilled over the sill from the outer basin, repressing the inner basin water to below 70 meters in the area closest to the sill.
• December
In December, stratification is noticeably weaker. Salinity is ∼34.5 psu in the upper layer and somewhat higher at depth. The fjord has cooled off further, and surface temperatures are∼-1◦C. A warmer layer (T'1◦C) is apparent just below 50 m depth.
In the outer basin, the cold bottom water has been replaced with warmer (T'1◦C) water.
• March-April (2003)
During spring, stratification is very weak. Salinity is almost homogenous (S'34.6 psu) with slightly higher values at depth.
The surface water is near or at its freezing point (T'-1.85◦C) while some heat is contained at depth (40-140 m) where temperatures up to -1.4◦C are observed.
The depth of the homogenous surface layer varies both temporally and spatially At the inner station, the layer deepens from about 80 meters on the 26th of March to 120 meters on the 11th of April. However, warmer intrusions (T'-1.4◦C) are found higher up in the water column. At the stations occupied on the 24-26th of April, the depth of the homogenous surface layer varies between 40 and 80 meters. In the shallow
”Petuniabukten” (station 14), the water column is homogenous from the bottom (50m) and up.
Freshwater content
The freshwater content presented in figure 5.1 was calculated using the bottom-salinity (34.8 psu) as a reference salinity and a ”mean” CTD-profile from the inner basin. The ”mean”
profile was obtained from the five CTD stations in section 1 that is situated inside the sill.
The total freshwater content decreases throughout the autumn, but shows a peak in mid-October. The peak can probably be explained by the warm period that Longyearbyen experienced in the beginning of October (see figure 4.11). For more than a week, Longyear- byen the warmest place in Norway and temperatures around 5◦C caused the snow to melt and the previously dry and frozen rivers in the Longyear- and Bjørn-valley to reappear and flood (personal observation). The situation around Billefjorden is likely to have been the
17/08 22/09 25/09 01/10 15/10 20/10 11/12 26/03 0
0.5 1 1.5 2 2.5 3
Freshwater content [m]
2002−2003
Total 60 m 30 m
Standard deviation
Figure 5.1: Freshwater content in Billefjorden, inner basin 2002-2003
5.2. DISCUSSION 29 Area Ice Thickness [cm] Salinity Fresh Water equivalent [m]
Ice Cover 1/3 80 7 0.2
2/3 40 7 0.2
Drifted ice 2/3 75* 7 0.3
Table 5.1: Estimated freshwater content in existing ice cover and the loss from drifted ice. *for discussion on thickness, see section 8.2.
same and the ”re-borne” rivers would then feed the fjord with freshwater and decrease the salinity. The warm temperatures would also increase glacier ablation.
The decrease in fresh water content takes place mainly above 30 meter, while greater depths experience a small increase. The increase at mid-depth can be explained by the deepening of the upper fresher layer and down-mixing while vertical diffusion is responsible for the increase at depth (see section 10.1.2).
Between December and March, the freshwater content is reduced by about a meter, from 1.5 to 0.5 m. Some freshwater is ”trapped” in the ice cover, while the rest has been advected either in liquid form, indicating that the estuarine circulation continued after 11 December, or in solid form, as ice drifting out of the fjord (about 2/3 of the ice cover drifted out of the fjord 5-22 February 2003, see section 4.5.1). Estimates of the freshwater contained in the ice-cover on 12 March 2003 and the loss from ice-drift is presented in table 5.1. Roughly 0.7 m of freshwater was estimated to be contained in or lost with ice, leaving circulation to account for another 0.3 m.
It should be noted, however, that the mean profile used to calculate the freshwater content in March was obtained from stations located in the inner, ice covered part of the fjord. These stations might not be representative since the total ice production probably is greater in the outer part, see section 9.2.
The loss of freshwater through ice-drift might be underestimated if the fjord acted as latent heat polynya at times when it was ice-free, see section 9.2.
Salt transport estimates from ADCP and CTD measurements
With relatively fresh water flowing out in the surface layer and more saline water entering at depth there is a net transport of salt into the fjord through the sill area. Balancing the velocity profile obtained from the ADCP in October 2002 (neglecting fresh water inflow to the fjord) by appointing a velocity of -12 cm/s to the surface layer (upper twelve meters), the salt transport can be estimated using equation A.2 in the appendix. A salt-transport of 8×103 kg/s was obtained using the velocity profile in figure 4.5 A and the salinity profile from figure 4.5 B.
The salt transport can be related to the observed decrease in freshwater content. The data from the 16 and 20 October reveals a decrease of 3mm/h or 0.3 m in four days, while the transport calculated above corresponds to a decrease of 7 mm/h. A freshwater input on the order of 100 m3/s is required to balance the numbers. As a comparison, the estimated freshwater input (360x106m3/year) correspond to a mean flux of 35 m3/s during the summer
months (June-September).
Heat Content
The heat content presented in figure 5.2 was calculated from the ”mean” profile discussed above. Heat capacity and freezing temperature were calculated from salinity and temperature following the Unesco (1983) routines (Matlab seawater package).
The available heat in the inner basin was steadily decreasing throughout the autumn, with the December value of 0.5 GJ being less than half of the 1.2 GJ contained in the fjord in August. The greatest reduction has taken place in the top thirty meters, where about 90%
of the heat has been lost to the atmosphere or through advection. The bottom layer, on the other hand, has gained heat through diffusion and mixing. The heat loss equals a negative flux of about 70 W/m2 throughout the autumn. Despite the observed warm inflow at depth, the maximum heat loss (∼180 W/m2) is found between the 15th and 20th of October.
The heat transport can be related to the heat content in the same way as the freshwater content was related to the salt transport. The heat transport was calculated using equation
17/08 22/09 25/09 01/10 15/10 20/10 11/12 26/03 0
0.2 0.4 0.6 0.8 1 1.2 1.4
Available Heat content [GJ/m 2 ]
2002−2003
Total 60 m 30 m
Standard Deviation
Figure 5.2: Available heat content in Billefjorden, inner basin 2002-2003
Flux [W/m2]
Advected heat, sill 100
Advected heat, fresh water -7
Sensible heat -100 to -200
Radiation, Latent heat ?
Observed heat loss -180
Table 5.2: Heat budget in Billefjorden, 18-20 October 2002
5.2. DISCUSSION 31 A.3 in the appendix and a heat transport of 10 GJ/s, correspond to a positive flux of 100 W/m2 over the fjord surface (inside of the transect), was found. The parameterizations for sensible heat flux described by Brown (1990) (equation A.4 in the appendix) and approximate meteorological values from Billefjorden during ”Mosby 4” (18-20 October: Tsurf = 2-4◦C, Tair = -3-5◦C and U10= 10 m/s) gives a negative sensible heat flux of about 100-200 W/m2 is . The sensible heat loss would thus roughly balance the advected heat but can not account for the large net heat loss (-180 W/m2) observed in the same period.
An input of cold (T'0◦C), fresh water from the surroundings on the order of 100 m3/s, as discussed above, and a corresponding outflow of warmer surface water (∼2◦) equals a negative flux of about 7W/m2. Latent heat and radiation (and computational/observational errors) accounts for the rest. The heat budget is summarized in table 5.2.
For cold water, like the water masses observed in Billefjorden, density is to a great extent a function of salinity while the effect of temperature is secondary. I.e. if ρ = ρ0+α(S − S0) +β(T −T0) then |α(S−S0)|| β(T −T0)|. We can thus expect greater variations in temperature in the horizontal plane than in salinity, and the uncertainties in the calculations are greater, see for example figure 4.4 B where temperature and salt profiles for five CTD stations are shown. Variations between stations are greater in the temperature profiles. The five stations used in the calculations of available heat might not be representative for the whole area.
Fjord Circulation
The circulation in a fjord is a result of external forces such as wind, tides and freshwater input and it is influenced by topography, friction and (if the fjord is wide) by rotation (Coriolis force). Fjord circulation is thoroughly discussed by e.g. Farmer & Freeland (1983).
Sill flow and tides will be discussed separately in chapter 7 and 8.
6.1 Theory
6.1.1 Estuarine circulation
Estuaries are defined by Cameron & Pritchard (1963) as ”semi-enclosed, coastal bodies of water which have a free connection with the open ocean and within which seawater is measur- ably diluted with freshwater derived from land drainage”. A fjord receiving fresh water runoff from the surroundings may thus be considered an estuary and an estuarine-like circulation is likely.
Aas (1983) explained how fresh water entering the fjord creates sloping surface (typically 1 cm per 10km, (Farmer & Freeland 1983)) and thus a pressure gradient driving the fresh surface layer seaward. On its way out of the fjord, the freshwater entrains and is mixed with the underlying water, and a deeper, brackish, (out-flowing) surface layer is formed. According to AMAP (1998), the outflow of brackish water might be 5-10 times as big as the fresh water input, and a compensating inflow is needed to replace the entrained seawater.
A schematic of the estuarine circulation is presented in figure 6.1.
Inverse Estuarine Circulation
If the fjord experiences a greater buoyancy loss than the water body outside (through e.g.
cooling, evaporation or brine-rejection), an inverse estuarine circulation can develop. The pressure gradient that builds up drives water out of the fjord at depth, while a compensating flow is established in the surface layer, see figure 6.2. Finnigan & Ivey (2000) accomplished this in laboratory experiments and showed that water from the stagnant bottom layer was entrained in the outgoing flow.
32
6.1. THEORY 33
Fresh Water input
Basin Water Intermediate layer Brackish layer
Entrainment
Sill
Figure 6.1: Estuarine circulation in a fjord with a freshwater source at the head.
Upper layerUpper layer Middle layer
Basin water Negative buoyancy flux
Sill
Figure 6.2: Inverse estuarine circulation caused by a negative buoyancy flux at the surface, from Finnigan and Ivey (2000)
The observed outflow of bottom water from Storfjorden (Quadfasel et al. 1988) during winter as well as the outflow through Gibraltar from the Mediterranean Sea are both examples of inverse estuarine circulation.
6.1.2 Rotational Effects
The Rossby radius, R, is the length scale at which rotational effects become important. If the scale of motion is larger than the Rossby radius, any movement will be deflected to the right(left) in the northern(southern) hemisphere.
In a homogenous fluid R is defined as
R=
√gH
f (6.1)
where g is gravity, H depth and f the Coriolis parameter.
In a two-layered system the internal or baroclinic Rossby radius, Ri is important. Ri is defined as
Ri =
√g0H0
f (6.2)
whereg0 = ρ2ρ−ρ1
2 g and H0 = HH1×H2
1+H2. ρ1 and ρ2 are the densities of the upper and lower layer respectively and H1, H2 are the thicknesses.
Thus Ri increases with increasing density difference and with increasing depth of the upper layer.
The Coriolis parameter, f, has an approximate value of 1.4×10−4 on Svalbard (78◦ N).
In a wide fjord where rotation is important the two-layered estuarine flow described above might develop in the horizontal plane. The out-flowing freshwater is deflected to the right and follows the righthand side of the fjord while the inflow is located on the lefthand side.
6.1.3 Wind effects
Wind blowing over water will exert a stress,τwind, on the surface. The stress can be expressed as
τwind =CDρairU2 (6.3)
whereCD is the drag coefficient, ρair is the density of air (approximately 1.20 kg/m3) and U is the wind speed at 10 meters height. CD is dependent on wind speed and on the stability of the air, see Pond & Pickard (1983) for discussion and section A.1 in the appendix for parametrization.
Due to the surrounding mountains, wind tends to be channelled along the fjord.
Narrow fjords
In a narrow fjord, where rotational effects are negligible, down-fjord winds accelerate the surface layer which is ”flushed” out of the fjord. Just as in the case of estuarine circulation described above, the outflow in the surface is balanced by a compensating flow at depth.
Wide fjords
If the fjord is very wide, i.e. if the width is greater than or equal to the Rossby radius discussed above, wind and rotational effects will interact and cause Ekman transport and up/down welling (due to friction and Coriolis the mean water transport is at right angle of the wind direction, see figure 6.3 B). The sides of the fjord will act as separated coasts with upwelling on one side and downwelling on the other. Geostrophic adjustment would eventually lead to geostrophic currents along the sides.
Cushman-Roisin, Asplin & Svendsen (1994) treated fjords as two-layer systems and used the concept of wind impulse to arrive at a criterion for when the surface layer can be ex- pected to outcrop, i.e for when the surface layer has been displaced to such an extent that the layer lower layer reaches the surface, see figure 6.3 A. When the wind impulse, I, defined in equation 6.5, is larger than Ioutcrop, defined in equation 6.4, the fjord can be expected to have outcropped.
Ioutcrop = f RH
H2tanh(L/2Ri) (6.4)