OR/14/015 South-west Deeps (West)

From Earthwise
Jump to: navigation, search
Lark R M. 2014. Mapping seabed sediments of the Swallow Sand and South-west Deeps (West) MCZs. British Geological Survey Internal Report, OR/14/015.

Data

The data used for the South-west Deeps (West) MCZ (SWDW) are particle size analyses from the verification survey carried out by Cefas in 2013 (McIlwaine, 2013[1]). The data were collected by 0.1 m2 Hamon grab from 208 locations on a pre-planned survey grid. Exploratory analysis found no duplicate locations in this data set. This is a sizeable data set to support a spatial analysis, so as described for Swallow Sand in section 2.1 above, no attempt was made to supplement these data with others from the BGS archive.

Exploratory data analysis and transformation

Exploratory statistics of the 208 particle size data are shown in Table 7 below. Eight zero values were recorded for gravel content and 2 for mud content. Non-zero values were imputed as for the Swallow Sand data, following Lark et al. (2012)[2]. Table 8 shows summary statistics of the alr-transformed data

Table 7 Summary statistics of particle size data from SWDW.
Gravel
/%
Mud
/%
Sand
/%
Mean 3.86 8.59 87.55
Minimum 0.00 0.00 42.85
Maximum 30.80 30.18 98.71
Standard
deviation

6.71

5.23

8.75


Table 8 Summary statistics of additive log-ratio transformed data from SWDW.
alr-mud alr-sand
Mean 2.01 4.55
Minimum 1.87 4.51
Standard
deviation

2.30

2.06
Skewness 0.035 0.448

Spatial analysis

The same procedures used with the Swallow Sand data were used to compute estimates of the auto- and cross variograms, and to fit LMCRs to the data for SWDW, and these are shown in Figure 6.

Figure 6 Estimated variograms and cross-variograms with fitted linear model of coregionalizations for alr-mud and alr-sand for SWSW. Solid symbols are method of moment estimates with fitted model in red. Open symbols are mimimum volume ellipsoid estimates with fitted model in blue.

Cross-validation results for the two sets of auto-variograms are presented in Table 9.

Table 9 Cross-validation results for SWDW.
Method of Moments me
Estimator
Variable
alr-mud alr-sand alr-mud alr-sand
Mean of θ (x) 1.17 1.02 1.85 1.51
Median of θ (x) 0.47 0.45 0.68 0.61

As for the Swallow Sand data the cross-validation results for the auto-variograms based on MoM estimates give results very close to those expected for the correct model and so the LMCR based on the MoM estimates can be used in further work. Its parameters are given in Table 10 below.

Table 10 Parameters of the linear model of coregionalization based on method-of-moments for SWDW.
Component Spatial correlation model type Distance parameter of spatial model/metres Variance or Covariance Component
Auto-variogram Cross-variogram
alr-mud alr-sand
1 Nugget N.A. 1.97 1.75 1.56
2 Spherical 6892 1.13 1.15 0.69
3 Spherical 28412 2.40 1.81 0.96

Spatial predictions

Spatial predictions were carried out as described for the Swallow Sand MCZ in section 2.3. The output files are SWDW_predictions.dat with the conditional expectations of gravel, sand and mud content and associated measures of uncertainty (same structure as described in Table 5 for the comparable file from Swallow Sand) and SWDW_classes.dat, with the EUNIS class probabilities (same structure as described in Table 6 for the comparable file from Swallow Sand).

Figure 7 below shows the conditional expectation of gravel, sand and mud across the SWDW MCZ, and Figure 8 shows the 0.025 and 0.975 quantile which define a 95% confidence interval for mud content. {{Clear}

Figure 7 Conditional Expectation of gravel, sand and mud content (proportions) across the SWDW MCZ based on multiple realizations of the prediction distribution.
Figure 8 Upper (top) and lower (bottom) bounds of the 95% confidence interval for mud content (proportion) across the SWDW MCZ based on multiple realizations of the prediction distribution .

Figure 9 below shows the most probable EUNIS class across the Swallow Sand MCZ, and the probability of the most probable class. Figure 10 shows the probability of each class.

Figure 9 Most probable EUNIS class (top) across the SWDW MCZ and probability of the most probable class (bottom).
Figure 10 Probability of finding each EUNIS class (top) across the SWDW MCZ.

As at Swallow Sand, class Sand and Muddy Sand is delineated as most probable across most of the SWDW MCZ. Uncertainty is greatest in central eastern parts of the zone, including areas where classes Mixed and Coarse have largest probability. While there is significant nugget variance in the LMCR for SWDW, there is stronger spatial dependence than at Swallow Sand. The mapped conditional expectations show local patches where the gravel content of the sediment is larger than average, and a general reduction in sand content of the sediment from south-west to north-east.

References

  1. MCILWAINE, P. 2013 South-West Deeps (West) recommended Marine Conservation Zone (rMCZ) Survey Report: C5785H. Issue date: 24 July 2013.
  2. LARK, R M, DOVE, D, GREEN, S, RICHARDSON, A E, STEWART, H, and STEVENSON, A. 2012. Spatial prediction of seabed sediment texture classes by cokriging from a legacy database of point observations. Sedimentary Geology, Vol. 281, 35–49.