Vicinity cell methods added to threshold outputs

[bayliffe]

Relates to: https://github.com/metoppv/mo-blue-team/issues/1020, metoppv/improver#2439

Threshold outputs with a vicinity applied updated to include an area maximum cell method.

Changes

For all but one of the diagnostic the only change is the addition of the new cell method, as shown using nccmp -dmNsgf, the output of which is below.

data/threshold/vicinity/kgo.nc

DIFFER : VARIABLE "probability_of_rainfall_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_rainfall_rate_in_vicinity_above_threshold : 4 <> 5

data/threshold/vicinity/kgo_collapsed.nc

DIFFER : VARIABLE "probability_of_rainfall_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo_collapsed.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_rainfall_rate_in_vicinity_above_threshold : 4 <> 5

data/threshold/vicinity/kgo_landmask.nc

DIFFER : VARIABLE "probability_of_rainfall_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo_landmask.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_rainfall_rate_in_vicinity_above_threshold : 4 <> 5

data/threshold/vicinity/kgo_landmask_collapsed.nc

DIFFER : VARIABLE "probability_of_rainfall_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo_landmask_collapsed.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_rainfall_rate_in_vicinity_above_threshold : 4 <> 5

data/threshold/vicinity/kgo_masked.nc

DIFFER : VARIABLE "probability_of_lwe_precipitation_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo_masked.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_lwe_precipitation_rate_in_vicinity_above_threshold : 4 <> 5

data/threshold/vicinity/kgo_multiple_vicinities.nc

DIFFER : VARIABLE "probability_of_rainfall_rate_in_vicinity_above_threshold" IS MISSING ATTRIBUTE WITH NAME "cell_methods" IN FILE "../reference/data/threshold/vicinity/kgo_multiple_vicinities.nc"
DIFFER : NUMBER OF ATTRIBUTES : VARIABLE : probability_of_rainfall_rate_in_vicinity_above_threshold : 4 <> 5

The only exception is the nowcast related test for which there were precision level differences in some values.

data/threshold/nowcast/kgo_masked.nc

DIFFER : LENGTHS : ATTRIBUTE : cell_methods : VARIABLE : probability_of_lwe_thickness_of_precipitation_amount_in_vicinity_above_threshold : 61 <> 71 : VALUES : "time: sum (comment: of lwe_thickness_of_precipitation_amount)" : "area: maximum time: sum (comment: of lwe_thickness_of_precipitation_amount)"

DIFFER : VARIABLE : probability_of_lwe_thickness_of_precipitation_amount_in_vicinity_above_threshold : POSITION : [1,0,46,73] : VALUES : 0.847467 <> 0.847467
DIFFER : VARIABLE : probability_of_lwe_thickness_of_precipitation_amount_in_vicinity_above_threshold : POSITION : [1,0,55,61] : VALUES : 0.62796 <> 0.62796
DIFFER : VARIABLE : probability_of_lwe_thickness_of_precipitation_amount_in_vicinity_above_threshold : POSITION : [1,0,59,65] : VALUES : 0.487856 <> 0.487856
DIFFER : VARIABLE : probability_of_lwe_thickness_of_precipitation_amount_in_vicinity_above_threshold : POSITION : [1,0,69,76] : VALUES : 0.357606 <> 0.357606
...

These same difference are found when comparing a new KGO generated using the master branch prior to this change, suggesting this is a result of some package change leading to precision level changes. These do not show in the standard running of the testing due to the allowed tolerances.

I will update this KGO in full as that better aligns the values with master's output as well as including the new cell method.

[Kat-90]

See improver PR for comment