Experiment Length Ambiguity

[charlotte]

Go through all the dates and make them consistent with one way of doing it, and be clear about how many years of simulation are expected.

[charlotte]

In experiment 6.5 cloud response I see:

<calendar>
    <daily-360>
        <range>
	    <closedDateRange/>
        </range>
    </daily-360>
</calendar>
<requiredDuration xsi:type="ClosedDateRange">
<!-- Idealised run so dates given here are only dummy values -->
    <startDate> 1850-01-01T00:00:00Z</startDate>
    <endDate> 1881-01-01T00:00:00Z</endDate>
</requiredDuration>

From that we don't know the duration of the experiment ...

By contrast in 1.1 decadal

        <shortName> 1.1 Decadal </shortName>
	<longName> 10-year prediction </longName>

	<calendar>
		<realCalendar/>
	</calendar>
	<requiredDuration xsi:type="ClosedDateRange">
		<startDate> 1959-09-01T00:00:00Z</startDate>
		<endDate> 2016-01-01T00:00:00Z </endDate>
		<lengthYears>56.25</lengthYears>
	</requiredDuration>

I fail to understand how we have 56 years of decadal experiments ...

Also, note the different way the calendar dates are handled.

Can one of you two please go through all the dates and make them consistent with one way of doing it, and clear about how many years of simulation are expected (obviously the first and last date might not help with that).

It's difficult with the decadal ones because there are a number of possible start dates, so you might have to come up with a way of encoding that variation, because those filling out the questionnaire may not do all the decadal runs, so they do need to be able to tell us which start dates and what durations they've run for ...

Meanwihle for the beta release, I'll leave these alone ...

Bryan

[charlotte]

This issue will be dealt with when we code up the decadal simulations as separate experiments. Handing over to Gerry because he is working on this.