Tuesday, December 16, 2014

December update on 2014 warmth.

NOAA has posted their report on November 2014 (h/t DK). It shows a global anomaly relative to 1901-2000 of 0.65°C. This is down from October's 0.74°C. TempLS grid was down 0.11°C, which looks like very good agreement, but is something of a break with the recent eerily close tracking.

They may take a few days to update the detailed MLOST file that I normally use (base 1961-90), so, with Christmas coming, I'll use a synthetic value by dropping the Oct value by 0.09°C too, so as to produce the last of the anticipatory posts about record prospects.

I'm following the format of an earlier post, with sequel here. You can click buttons to rotate through datasets (HAD 4, GISS, NOAA, TempLS mesh and grid, and HADSST3). I haven't shown the satellite troposphere indices, because these are nowhere near a record. The plot is cumulative sums of monthly ave relative to 2010, the next highest. I see that NOAA has a similar plot, but with average to date rather than total. This is just a scale difference, becoming small near end year.

Highlights are, first HADSST3, which is way ahead of 2010. In fact, for that, 1998 was higher, at 0.43°C, but that level too should be exceeded. This emphasises that high SST was the driver for 2014.

GISS is less clear; Nov 2014 was cooler while Nov 2010 was warm, so it's prospects receded slightly. Meanwhile, my NCEP/NCAR daily index showed the first week or so in December very cool, but then warmer. So GISS is no certainty. However, Dec 2010 was quite cool. NOAA is well ahead, and while there is no November data for HADCRUT 4, it is also well placed.

Meanwhile, the The plot is below the jump:

Sunday, December 14, 2014

November was cooler - GISS down 0.11°C

GISS has reported for November 2014, down from 0.76°C to 0.65°C. The title alludes to my post of 4 Dec, which reported a similar drop by TempLS Mesh. That drop has reduced somewhat as more data arrived. I have been following these monthly events with added interest, because it is the first real test of early TempLS mesh predictions (described here), and was foretold by another index I have been calculating (reanalysis NCEP/NCAR). This reports daily, and indicated a considerable cooling in November.

As usual, I have also done a TempLS Grid calc. As explained here, I expect the mesh calc to more closely emulate GISS, while the grid calc should be closer to the NOAA and HADCRUT indices. TempLS grid did show a greater reduction, from 0.631°C to 0.519°C.

When NOAA comes out, I'll post on the likelihood of a record this year. The NCEP/NCAR index has been quite cool in December as well, so a GISS record is in the balance. NOAA and HADCRUT are currently ahead by a greater margin, but TempLS suggests may see a greater reduction.

Details of the GISS map for the month, and comparison with TempLS, are below the jump.

Thursday, December 4, 2014

November was cooler - latest TempLS

Since I wrote in October about a new scheme for early mesh-based TempLS reporting, I've been looking forward to producing such an early result. Last month didn't work, because GHCN was late. But this month everything is on time. For me, there is added interest, because I developed the daily NCEP/NCAR based index, and it has been suggesting a perhaps unexpected drop in November temperature.

The early TempLS mesh report is now out, and it does show a corresponding drop, from 0.647°C (Oct) to 0.557°C (Nov, base 1961-1990). The Oct value also came back a little, which reduced the difference slightly.

It is indeed a very early report, and will change. 3108 stations, probably about 70% of final. There is essentially no data from Canada, China, Australia, and most of S America and Africa. So it's too early to have much faith in the regional map, but the cold in the US certainly showed up.

Both satellite indices showed small reductions, with RSS going from 0.274°C to 0.246°C, and UAH similar. In terms of a record warm 2014, I think the likelihood is essentially unchanged.

Update 8/12 With most GHCN data now in, TempLS has risen a bit, to 0.579°C. But the NCEP/NCAR measure went the other way. It stayed cold in November, with the average down to 0.106°C, and December (to 4th) has been cold too.

Wednesday, December 3, 2014

Reanalysis revisited

I have been working with climate reanalysis. I found some more resources, mainly through the Univ Maine Climate Change Institute. They have a collection of reanalysis offerings, some of which are just remappings of flat plots on the globe. But they have useful information collections, and also guides. A comprehensive guide page is here. They have a collection of GHCN daily data here, convenient, although not up to date. They have globe maps of daily temperature, as I do here, but again with a considerable lag. And they have a section on monthly reanalysis time series here, which is the focus of this post.

They let you draw graphs of annual data, and plots of each month over years, but frustratingly, not a monthly plot. This may be to avoid including seasonal variation; they are not anomalies. However, they do give tables of the monthly average temperature for various reanalysis methods, to only 1 decimal precision :(. Despite that limitation, it is useful for me, because I had wondered whether the convenience and currentness of the NCEP/NCAR data was undermined by its status as a first generation product. I now think not; the integrated global temperature anomaly is quite similar to more advanced products (MERRA is something of an outlier). More below the fold.

Monday, December 1, 2014

Maintained monthly active temperature plotter

This post follows on from a thought by commenter JCH. On the latest data page, I maintain an active graph of six recent temperature indices, set to a common anomaly base 1981-2010. But I actually maintain a file of about fifteen. JCH mentioned a difficulty of now getting recent data for HADCRUT, for example. So I thought I should add some user facilities to that active graph to make use of this data.

I tried dynamic plotting once previously with annual data - it is the climate plotter page. It hasn't been much used, and I think I may have tried to cram too much into a small space. So I've been experimenting with different systems. I've learnt more about Javascript since then.

So the first addition is a panel for choosing which datasets to show. It has a floating legend, with buttons for changing color, asking for regression (OLS), or smoothing (12-month boxcar). If you ask for regression, a similar panel pops up with again color, and start and end time text boxes. Initially it sets these to the visible screen, but you can type in other times. If you press on any color button, another panel pops up with color choices. The OLS trend in °C/Century for the stated interval, is in the red-lined box.

The original plot worked by mouse dragging. If you drag in the main space, the plot follows. But if you drag below or left of the axes, it stretches/shrinks along that axis. I've kept that, but added an alternative.

There is a faint line at about 45° from the origin. If you move the mouse in that region, you'll see faint numbers at each end of the x-axis. These are tentative years. If you click with Shift pressed, the plot adapts so that those become the endpoints. The scheme is similar to the triangle of the trend viewer, but backwards. Near the origin, you get short intervals in recent time; the scale of mouse move gives better resolution here. Move along the x-axis makes the start recede; along the diagonal, both recede keeping the interval short. In the upper triangle, it's similar with the y-limits. It's easier to try than to read.

The Redraw button is hardly needed, because there is much automatic redraw. The Regress button forces a recalc when you've manually edited the text boxes for intervals. Each pop-up window has an exit button; the Legend button is the way to bring it back (it toggles). Each pop-up is draggable (gently).

So here it is below the fold. It's still experimental, and feedback welcomed. When stable, I'll embed it in the page.
Update - I seem to be using JS things not supported in Chrome or IE. Under construction. Everything  seems OK in Firefox. But the plot dragging still doesn't work in Chrome or IE.
Seems to be fixed now.

Sunday, November 30, 2014

Cooler November?

The main purpose of this post is to note that the daily NCEP data is now regularly updated here. As with TempLS mesh, there is a kind of Moyhu effect whereby when I set up a system like this and want to tell everyone about it, there is a hiatus in the data source. This time I think it is just Thanksgiving.

Anyway, the global story it tells is that there was a cool dip around Nov 13, at the height of the N America freeze, and a second a few days later. Currently (to Nov 24) the average anomaly for Nov is 0.157°C, compared with 0.281°C in October. I think this will pan out to November being about 0.1°C cooler than October in the surface temperature indices.

What does this mean for talk of a surface record 2014? I think it is neutral. To reach a record, month temperatures have to exceed on average the 2010 average, and it looks like November will be close to that number. For example, GISS Oct was 0.76; 2010 average 0.66. This may matter for GISS, which was only just above the average anyway to date. NOAA and HADCRUT have a greater margin.

Update: Three more days data arrived, and still cool. The month average is now down to 0.135°C, a drop of about 0.15 from October. That is negative for GISS record prospects.

Tuesday, November 25, 2014

Daily reanalysis NCEP/NCAR temperatures with WebGL.

In a previous post I described how a global index could be created simply by integrating the surface temperature data provided by NCEP/NCAR. This data is to within last few days, and I've described here how numerical data is maintained on the latest data page.

As gridded data, I then sought to display the daily temperature anomalies (base 1994-2013) with WebGL, and that is shown here. I'm also planning to maintain this, on the latest data page if it does not drag out loading. Currently the daily data is just for 2014, although that is easily extended.

So it's below the fold. As usual, the Earth is a trackball that you can drag, zoom (right mouse) and orient (button). I'm trying a new way of choosing dates. High on right there is a tableau of small squares, each representing one day. Click on this to choose. It's all a bit small, but to the left of the Orient button, you'll see printed the date your mouse is on. So just move until the right day shows, and click.

Because temperature ranges are large, it has been quite hard to get the colors right. You might like to look at the recent North American cold spell for shades of blue. Incidentally, I see that the global average has slipped again, so November looks like a much cooler month than recently.