SIO Upper Ocean Temperature climatology and Anomalies

The climatologies include temperature, heat storage, and mixed-layer depth, while the anomalies include monthly heat storage. They are served as datasets, including an interactive viewer and downloadable data files.

Mixed-layer depth is defined as SST - 1.0C.

Annual temperature profile observation counts
YEAR	COUNT
1955	41,958	NODC NGOTS
1956	45,978
1957	57,427
1958	67,380
1959	65,886
1960	68,758
1961	76,295
1962	82,125
1963	89,042
1964	83,430
1965	85,909
1966	93,860
1967	74,938
1968	74,305
1969	68,484
1970	59,781
1971	65,320
1972	60,205
1973	67,360
1974	71,294
1975	59,659
1976	60,606
1977	60,171
1978	61,217
1979	63,189
1980	63,715
1981	56,893
1982	54,920
1983	48,150
1984	50,863
1985	50,294
1986	49,072
1987	42,969
1988	26,260
1989	40,092	GTSPP
1990	13,039
1991	21,790
1992	29,515
1993	54,248
1994	51,520

TEMPERATURE CLIMATOLOGY AND ANOMALY ANALYSIS PROCEDURE

Temperature profile databases used:
1. 1955 to 1988: NODC Global Ocean Temperature/Salinity CD-ROM data set.
2. 1989 to 1994: Global Temperature/Salinity Pilot Project(GTSPP) data set.
Interpolate temperature profiles at inflection points to 15 standard levels(0, 20, 40, 60, 80, 120, 160, 200, 240, 300, 400, 500, 600, 700 and 800 meters). Simple linear interpolation is used.

CLIMATOLOGY

Calculate MONTHLY climatology at each standard level by BIN AVERAGING absolute temperature profiles to a 5 Deg.(X) -by- 2 Deg.(Y) grid res- olution. Only data from Jan. 1980 to Dec. 1989 was used. The MONTHLY mean was calculated for each month from 1980 to 1989. Each monthly mean is then summed over the 10 year period to form the RAW climatological mean for the month.
A Fourier analysis is performed using the first 2 harmonics(the ANNUAL and BI-ANNUAL cycles) for each time sequence at each grid location for each standard depth. The Fourier series is
```
     y'(i) =  [An * COS (2*n*pi*t/12) + Bn * SIN (2*n*pi*t/12)]
          for n = 1 to  2 and
              t = 0 to 11
```
The Fourier coefficients A1/B1, A2/B2 and the long-term ANNUAL mean are written-out as grids at each standard level.
Areas of missing data in the Fourier coefficient and long-term ANNUAL mean fields are filled by zonal and meridional linear interpolation. The following procedures are followed for the area filling,
1. The maximum ZONAL no-data gap is 20 degrees.
2. The maximum MERIDIONAL no-data gap is 10 degrees.
New monthly climatology fields are constructed using the filled Fourier coefficients and long-term ANNUAL mean.

ANOMALIES

Temperature anomaly profiles are calculated using the reconstructed monthly climatology. The climatology is BI-LINEARLY interpolated to the location of each temperature profile and the anomalies calculated at each standard level.
Profiles of ABSOLUTE, ANOMALOUS and CLIMATOLOGICAL temperature are saved in MONTHLY files.
Monthly Optimum Interpolation mapping of the temperature anomalies is performed for each standard level(0 ==> 400M) for each month from Jan. 1955 to Dec. 1994.
Decorrelation scales used for th O.I. mapping:
```
      ZONAL(Lx)      = 10 degrees
      MERIDIONAL(Ly) =  5 degrees
      TEMPORAL(Lt)   = 90 days(from the 15th of each month).
```
The O.I. mapped fields are filled in space and time. The spatial filling procedure is performed first with the following limits,
1. The maximum ZONAL no-data gap is 20 degrees.
2. The maximum MERIDIONAL no-data gap is 10 degrees.
The temporal filling is done with Maximum Entropy Prediction based on the algorithm by N. Andersen, 'On the Calculation of Filter Coefficients for Maximum Entropy Spectral Analysis', 1974, Geophysics, Vol.39, No. 1, pp69-72.
Adjust the space/time filled mapped monthly anomalies by computing the monthly MEANS of the ANOMALIES(over the period Jan. 1955 to Dec. 1994) then subtracting these means(of the anomalies) from the mapped fields.

Reference

The original dataset is kept here. The original dataset was produced by Ted Walker (walker@canopus.ucsd.edu).