TABLE 3.

Network structure corresponding to type of input layer, shape of the output from each layer, and the number of parameters associated with each layer of the CNN model which takes SSTA as input.

Layer (type)	Output shape	Parameters
reshape (Reshape)	(None, 46, 84)	0
lambda (Lambda)	(None, 46, 84, 1)	0
conv2d (Conv2D)	(None, 46, 84, 64)	320
average_pooling2d	(None, 23, 42, 64)	0
(AveragePooling2D)
conv2d_1 (Conv2D)	(None, 23, 42, 16)	4112
average_pooling2d_1	(None, 11, 21, 16)	0
(AveragePooling2D)
flatten (Flatten)	(None, 3696)	0
dropout (Dropout)	(None, 3696)	0
dense (Dense)	(None, 2250)	8,318,250
dropout_1 (Dropout)	(None, 2250)	0
dense_1 (Dense)	(None, 1125)	2,532,375