Scottish Marine and Freshwater Science Volume 4 Number 1: Annual Cycles of Physical, Chemical and Biological Parameters in Scottish Waters

Methods

1. Source Data

The source data from which the database has been constructed consist of computer files in two standard formats; .HYD format containing depth, temperature and salinity data and .CEM format containing standard chemical and biological parameters. A .HYD file exists for every research vessel cruise carried out since 1960 and a .CEM file if there were chemical measurements performed during the cruise. In addition, each station within the file contains header information such as latitude and longitude of the station, time and date, sounding and the number of depths sampled at the station. The data in the .HYD file can either have been collected by reversing water bottle or by profiling CTD instrumentation. Where CTD instruments have been used the .HYD file contains a condensed form of the data collected. This means that no more than twenty one depths are extracted from a CTD instrument profile for inclusion into a .HYD file.

2. Database

In order to facilitate rapid data searching and retrieval, the standard format files have been processed using an intermediate program which results in a large disc-based file amalgamating the data from many individual cruises.

For two previous reports (1992 and 1999) the database files were stored as Record Management Services ( RMS) files. This type of database structure became redundant due to the decommissioning of our VAX mainframe system.

For the last report (2005) the above database was converted to a flat ASCII file. This file now contains over 318500 lines of depth data up to the end of 2010 on which this report is based. Each line of data contains the following information: position (latitude and longitude), station number, cruise number, date (day, month, year), time ( GMT), International Council for the Exploration of the Sea ( ICES) ship code, sounding and pressure and one or more of the following parameters: temperature, salinity, density, oxygen, saturated oxygen, phosphate, nitrate, silicate, carbon, nitrogen, chlorophyll and phaeophytn.

The table below provides a summary of the data held in the database file:

Five year period	Total number of stations	Total number of individual depths
1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-09 2010	4,281 2,391 3,705 1,948 1,759 2,853 3,384 3,808 5,775 5,709 501	18,448 10,198 16,844 10,905 9,828 21,535 35,059 40,535 74,118 73,613 7,499
Total	36,114	318,582

The database file occupies approximately 56.6 megabytes of disk space. At each depth, as many as eleven individual parameters may be measured. The increased use of CTD profiling instruments can be detected from the table above. A dramatic increase in the number of individual depths used in this report can be seen from the late 1980s to the present date. This can largely be put down to the increased use of CTD's. Only standard depths are extracted from the CTD data files for inclusion in the data base. The number of standard depths extracted from the CTD data files never exceeds twenty one and the standard depths are determined by the sounding listed on the station.

In this paper twenty six sub-areas have been defined geographically and represent distinct hydrographic regimes. Monthly means and standard deviations have been extracted for all of the standard parameters regularly reported in the .HYD and .CEM files. From these means and standard deviations annual cycles have been computed (see below). A suite of programs has been written to search the database file to extract the data and to produce these tables of means and standard deviations for the twenty six sub-areas listed below.

3. Measurement Techniques

The tables present the monthly mean values, where data is available, of the eleven observed parameters and of two derived parameters (density and stability). The various methods used to determine these values are briefly described:

Temperature (T): This has traditionally been measured using reversing thermometers mounted on Knudsen bottles. Thermometers are held at set depths for six minutes to allow equalisation to the ambient temperature before the bottles are triggered. Increasingly since the late 1970s temperatures have been obtained using temperature sensors mounted on CTDs. These and the reversing thermometer readings are cross-checked against each other to identify erroneous measurements or instrument faults. After the mid nineties approximately 95% of all temperature readings were taken using CTDs, the reliability of these readings determined by manufacturer's biannual recalibration. Accuracy = +/-0.02°C.

Salinity (S): The traditional method employed over the period 1960-1990 has been to collect water samples at standard depths using Knudsen reversing water bottles. Salinity is then determined using a bench-top inductive salinometer, calibrated using IAPSO standards. Since about 1980, salinities have been increasingly determined in situ using conductivity sensors mounted on CTDs. These are calibrated against samples collected by reversing bottles at calibration stations performed periodically throughout a cruise. Accuracy = +/-0.005. As salinity is now defined as a ratio no units are used ( UNESCO, 1978).

Density (SIG-T): The density anomaly has been computed using the formulae presented in UNESCO (1983) with the required input parameters of temperature and salinity. Accuracy = +/-0.01 kg m ^-3.

Oxidised nitrogen (NO ₃): Samples for the determination of inorganic nitrogen are stored in a deep freeze. Concentrations are determined as nitrite following reduction with copper-coated cadmium (Stearns and Strickland, 1967). Accuracy = +/-10%. Units = μg-at/l.

Orthophosphate (PO ₄): If samples are not to be analysed immediately, the samples are fixed with chloroform and stored in a cool place. Concentrations are determined by a manual phospho-molybdate complex colorimetric method (Murphy and Riley 1962). Accuracy = +/-10%. Units = μg-at/l.

Silicate (Si): Samples are stored in plastic bottles in a cool place. Concentrations are determined by a silicomolybdate manual colorimetric method (Mullin and Riley 1955), later modified following Strickland and Parsons (1968). Accuracy = +/-10%. Units = μg-at/l.

Ammonia (NH ₃): Samples to be used for ammonia determination are stored frozen. A manual colorometric method is used, using phenol hypochlorite with potassium ferrocyanide to form indophenol blue complex. Accuracy = +/-10%. Units = μg-at/l.

Particulate Organic Carbon ( POC): Samples are filtered onto glass fibre filters (Whatman GFF) and the filters deep frozen. Analysis used to be done by a wet oxidation method, followed by titration with a thiosulphate solution (Strickland and Parsons, 1968). Since the early 1970s all analysis has done by a combustion method with measurement on a Perkin-Elmer CHN analyser.

Thawed samples are exposed to hydrochloric acid fumes prior to combustion in order to remove inorganic carbonate. Accuracy = +/-10%. Units = μg/l.

Particulate Organic Nitrogen ( PON): These values are obtained from the Perkin-Elmer analyser using the same sample as used to determine particulate organic carbon. Accuracy = +/-10%. Units = μg/l.

Chlorophyll- a ( CHL): Samples are filtered onto glass-fibre filter papers and stored in deep freeze. The chlorophyll- a is extracted by grinding the filter papers in a solution of 90% acetone and determining the concentration on a flourometer. Accuracy = +/-10%. Units = μg/l.

Phaeophytin ( PHAE): After the chlorophyll- a concentration has been determined the samples are dosed with dilute hydrochloric acid and the phaeophytin concentration again determined flourometrically (Strickland and Parsons, 1968). Accuracy= +/-10%. Units = μg/l.

Oxygen (O ₂): At the time of collecting the samples, the oxygen is fixed with manganese chloride and alkali-iodide solutions. Dissolved oxygen concentrations are determined using the standard Winkler method. These are converted into percentage saturations using the recorded in situ temperature and salinity. Accuracy = +/-10%. Units = % saturation.

Stability ( STAB): The stability of the water column is expressed as the log of the mean of the potential energy anomalies (Simpson et al., 1978) calculated for each individual station in a sea area. The potential energy anomaly is defined as:

4. Data Processing

Search methods

As described above the main database has been searched using boxes whose boundaries are defined as latitudes and longitudes or as polygons whose vertices are also defined as latitudes and longitudes. The boundaries of each box and the vertices of each polygon are presented as header information in each summary data table, and are summarised in Figures 1- 5.

In previous Annual Cycles Reports (1992, 1999) the data from the areas (Figures 4-5) defined along the standard hydrographic sections ( JONSIS, Fair Isle/Munken and Nolso/Flugga) were extracted as groups of individual stations. For example, the Fair Isle Current ( FIC) area contained only data from the JONSIS stations 1, 1a and 2. In Annual Cycles Report (2005) the boxes were defined to extract the data for the areas FIC and ONNS and polygons are defined to extract the data for areas AW_MUNKEN, AW_NOLSO, FCCS_MUNKEN, FCCS_NOLSO, FCCN_MUNKEN and FCCN_NOLSO. This extraction method is also used for this report. The change in the extraction procedures for these areas results in the inclusion of station data that were sampled at other times other than when the standard sections were being sampled.

In addition, the areas defined along the standard hydrographic sections (The JONSIS, Fair Isle/Munken, and Nolso/Flugga sections - Figures. 4 and 5) are representative of different water types. The indices so derived have been used previously by Turrell (1992a, 1992b).

Calculations

Monthly means and standard deviations within each of these areas, and between set depths are also presented. If no data is available in any month, a value is derived by linear interpolation using adjacent monthly mean values. If a gap of more than three months exists, then no interpolation is performed and no annual cycle is derived. The number of data points used to compute each monthly mean is also archived.

Annual cycles of each parameter, A, have been calculated using a least -squares fitting procedure, of the form:

where A _avg = annual average value, A _amp = amplitude of annual cycle, t = time in months (1st January = 0), Φ = phase shift (months) and Y = 12 (Maddock and Pingree, 1982).

It should be noted that while a sinusoidal annual cycle may describe the variability of parameters such as temperature and salinity accurately, such functions are not always suitable for chemical or biological parameters. Although the program which generates the tabular output automatically fits sine functions, and these may be useful in a variety of applications, the mean values must be closely inspected to determine if such a function is valid

The depth limits for the derivation of surface and bottom values have been determined from previous work in each sub area and from a preliminary examination of the data.