National Climate Database open for free access! :)

[RWood]

Hmm I see that “wifey” decided to be naughty whilst I was answering the call of nature and inserted the “doctor and nurse & middle aged” & “we are confused and where we should go” comments to embarrass me, the fact that she had 2 bottles of wine may have contributed largely….. I didn’t even notice that part when previewing & I seem to have forgotten to edit/finish it off, but I was also very tired and when I am very tired my eyes don’t work well - I all but have to hold them open with matchsticks


I mentioned that Ob Hill started recording sun from 1992 - (this part was not finished/edited & with the silly last comment from “wifey”) - what I meant to say was that in the M2 bulletin - sunshine hours were used from Ob Hill until 1992, then figures from the airport were used (and still is nowadays). The change of sites is because of obstructions from tall buildings in the city which particularly in latter years had an effect on sunshine hours .

Yes - obstructions can be an issue. That's why I think reading from airports are often a good idea as they are typically measured in locations with low horizon obstructions, or at least not blatantly obvious ones close by, like vegetation or very close buildings. See also my reply about the % calculations.

PS - My ever-loving partner is by now well used to my weather- (or primarily climate-) obsessed ways. She just expects me to give infallible local forecasts as well!

[RWood]

My replies below in italics:

Invercargill: I've seen old notes about the site that certainly state that there were observational issues near the end of its life, at least.

But setting that aside, the Nov 1925 might be OK. - but a variety of similarly unpleasant Novembers has never produced a total less than 129 at the Airport site. Also, the Dunedin total for the month was 155. So the In'gill total was only 66% of the Dunedin one - for the "new" site, that ratio has never been less than about 76% for any of the months Oct-March inclusive. The E-SE profile for March 1915 doesn't really accord with a low value at Invercargill (Dunedin would be different). If it were westerly dominated I think it would be more plausible. And Dunedin recorded 125 hrs for that month.

My overall impression is that the site was probably inferior, certainly in its later years, and that those 2 totals at least are pretty dubious regardless. Something to ponder.

One thing is certain: the June 1919 total quoted as 173.8 is clearly a typo. I'm sure that value was never in the old paper reports.[/quote]

Right I see, re November's sunshine total - yes you could be right, but i think its something we'll have to ponder for a long time unless one of us can ask NIWA about this.

I think Nov. 1925 will have to remain something of an unknown - unless by some miracle NIWA has old source data on paper and some detailed daily reports.

I can see the June error, I think it should be reading as 73.8 hrs as we both know there (unless by a very huge miracle) is no way Invercargill can record 173 hrs for a June.

Agreed! And I think when I saw stuff in Statistics books that that number was either missing, or had a sensible value like 73.8.
February 1919 (85 hrs) and 1926 (89 hrs) are pretty dubious too, because not once at the Aero has a Feb. total of below 100 hrs been recorded.

Now I dont know if you have noticed but take a closer look at the sunshine totals for sites 5812 and 5814 - those from 1932 to 1951 show the same hours althought from different sites, is this in error? & the city site didn't record any more sunshine since the move to the Aero in 1932.

They have just inserted the new Aero values into the other record as well, for some reason. The old site definitely didn't record that data once the Aero took over.
Where is Arapito? I cant find it on the map, is that in Westland?[/quote]

Arapito is about 10km inland from Karamea, in the far north of Westland. Interestingly, I thought its sun values from the start in 1979 to the end of the 80s looked significantly too low. Sure enough, from about 1990 onward they increased considerably, both in absolute terms and by comparison with Greymouth and Hokitika. Some improvement in exposure or other fix of a problem at the site must have done about then.
BTW, I have set myself the task of getting all of the 1971-2000 rainfall normals (about a 1000 of them!) from the database. I had already paid for 160-odd a couple of years ago - bugger!

PS - I also have the older Dunedin records (Jan 1930 and earlier - resumption after a gap was in 1933). They look pretty good, with the possible exception of a value of 78 for March 1922. I'd have to see a summary description of that month to get a bettere feel for it.

[janewaystv]

Looking at March's 1922 summary - it was a stormy month of frequent W disturbances with western areas getting an excess of rain above the average whilst eastern areas had an deficit with Gisborne/Napier recording less than half the average with floods on the SI's WC and some severe gales were experienced on various days throughout the month with a tornado at Wellsford (Auckland District) on the 26th causing considerable damage to property.

IVC for that month had 106 hrs, DUD is normally sheltered from W weather so I'm inclined to think the record of 78 hrs might be in error given that it was a month of W weather.

[RWood]

Ah yes - Wellington's MSLP reading for March 1922 was only 1003.7 mb - a full 6.5 lower than the next lowest! (records 1911-present). That 78 for Dunedin looks very fishy indeed in light of the report you've quoted.

[janewaystv]

by RWood on 3/10/06

If you look at the data type request panel, there is a selection of daily options available, as well as monthly obs. or means.
I'll give more detail tomorrow if you still can't find it - going to bed now.

Edit: Daily/hourly obs. are the first choice in the datatype selection menu (not neceassarily as complete as monthly summaries as not all the old data may have been kept in detail. For example, Kelburn sunshine monthly values go back to Jan 1928 but the available "dailies" start at Jan 1954.

Hi, have copied & pasted your above post from the Sydney's weather thread to this thread.

Have finally done some downloading of dailies (sun, rain & min/max temps) for IVC (starting with the year I was born - 1976 - will do pre 1976 at a later time), I did a test run & discovered if you enter this format = 01/01/76 to 31/12/76 then the data for the 31st is omitted so best to use this format = 01/01/1976 to 01/01/1977 - the data for the 31st is included.

Interesting reading so far

[RWood]

Yes, you have to watch the dates. The 0900 readings for a given date credit the max. temp. and rainfall to the previous day, but the min. temp to that day. Sunshine of course is also for the previous day.

I was older in 1976 than you are now - must be old!

[RWood]

Note for janewaystv: I am downloading many monthly values that I already had on record from the old Met.Observations periodicals. This has given me a chance to compare sets of data. In some cases monthly values that were present in the print publication are missing - I suspect that estimates were done at the time, but these have not been retained. I intend to keep those values whenever they look reasonable. Occasionally monthly values not present in Met. Obs. are in the downloaded versions - again, I'm keeping them if they look reasonable. Where 2 values disagree I accept the download version if it looks OK.

[RWood]

Still happily downloading records. It's a bit frustrating that so many places have had a change of site so truly longterm rainfall records are not very common. Queenstown is good - 1871 onwards, with just some years in the 1880s missing (I suspect that happened to several sites around then and it's possible govt. policy had something to do with it). Ashburton has a continuous record from 1909 with just 1 missing month. Many others to check as yet.

[janewaystv]

Tell me about it, I've experienced the same when downloading data for coastal Southland stations as well as the 4 main centres - one thing I really noticed for several sites post -1990 - many sites have missing data in between months/years, not to mention the numerous changes in sites. - those actions does make it a bit hard to get "proper" averages so I've consolidated several sites (for the same town/city) datasets into one. I've finally downloaded all daily sun, rain & temp obs for IVC Aero - filled 4 marbig folders to the point of bursting , also done with several Auckland sites for rain.

I've compared the downloaded data with that of my grandfather's data - there are quite some differences, I think I would go with the recent downloaded ones as the records from my GF were probably not verified/checked at the time he got them.

Coastal Southland stations I've gathered data for are Bluff (all sites), Stewart Island (all sites), Tiwai Point and remaining IVC ones, for those areas - widespread "dry" years that stand out of current cals. I've done are 1943, 1947 and 1956 - outstanding for the Stirling Point site - only 659mm - much less than IVC's record lows of 780-830mm set in 1947.

Bluff has recorded rain at 3 different sites at various times, East Bluff (SP) seems to have higher rainfalls then Western Bluff (R) (based on averages I've done so far), I find it a bit odd as SP is on the sheltered side from the Westerlies & R on the more exposed W side, probably the Hill does influence rainfalls to some degree around Bluff as well as giving some shelter, I remember speaking to a local ages ago, he said the Hill causes showers to deflect to either side of Bluff - that is to say showers go across Invercargill more frequently which might explain the rather lower totals for Bluff.

[RWood]

Yes, interesting stuff indeed. Assessing local influences on rainfall can be very tricky, so site changes are always a bit annoying. At least when there's an overlap of a few years one can do a bit of possible adjustment if there's a significant difference.