I looked up the Berkeley data. 1911-1930 average: -12.6C.  No data before that.  Berkeley also has -9.9 for 1931-1960, but -13.3 for 1961-1990.  The warmest December on record was in 1938: -1.C.  The coldest in 1988: -22.4.

Looking up the annual temperature trends for Svalbard on Berkeley Earth, I actually get a record back to 1761:
svalbard-and-jan-mayen-TAVG-Trend.png

That is remarkable given that there were no systematic observations before 1911, and only very few scattered observations before that.  Indeed, prior to 1911 all stations are more than 1000 km outside the area:
svalbard-and-jan-mayen-TAVG-Counts.png

The graph shows a stable climate till around 1910, then there is a dramatic change and the climate suddenly began to oscillate.  1910 also marks the year when actual in situ measurements began.  Are we to believe that the stability prior to 1910 is the correct picture and the oscillation after 1910 is the signature of anthropogenic influence?  Frankly, I think the reconstruction prior to 1910 is utterly junk.  The 95% uncertainty range - seriously? 

I don't understand why researchers make such bold statements when there clearly is little data.

The fact that trends "happen" to change when we begin to measure something accurately is, of course, not a coincidence every time.  We see the same thing for arctic sea ice coverage.  Reconstructions show a steady ice cap until a decline since 1979, which also happens to be when satellite records start.  Even worse, 1979 coincides with the low of the last AMO cycle (assuming that the AMO is not a fantasy).  I've said it many times over the past decade: Don't be surprised if it turns out that the sea ice increases after 2020 (till the 2040's).  AMO probably turned again around 2015.

What you presented is very interesting- do you have access to New York City climate data? 

the compulsive thinking in climate science: "If obsvervations don't exist, there must be a way somehow to reconstruct them".

Sometimes the answers are "no" and that line of research hits a dead end, but surprisingly often and with a bit of ingenuity they can be "yes", which is what makes paleoclimate science possible.

What you presented is very interesting- do you have access to New York City climate data? 

You can look up Berkeley Earth's New York Data. Their data before 1890, certainly before 1850, is pretty worthless in my opinion, reflecting the compulsive thinking in climate science: "If obsvervations don't exist, there must be a way somehow to reconstruct them".

Do also look at data for individual stations.

the compulsive thinking in climate science: "If obsvervations don't exist, there must be a way somehow to reconstruct them".

The actual and reasonable thinking is: "If direct observations do not exist, then do there exist methods for reconstructing them by correlations to other observables, are the mechanisms behind those correlations testable, and do independent methods produce results that are consistent and sensible?"

Sometimes the answers are "no" and that line of research hits a dead end, but surprisingly often and with a bit of ingenuity they can be "yes", which is what makes paleoclimate science possible.  A good example is reconstruction of surface temperatures across Antarctica.  Obviously there are few direct observations before the satellite era, and ice core sites are few and far between.  But Antarctica has a lot of topography, and there is a powerful correlation between altitude and temperature (the lapse rate), which among others (e.g. with latitude and distance from the coast) helps produce an interpolation method for temperature between sites.  An important area of paleoclimate research is the formulation and testing of models that connect sparse and localized data to regional and continuous climate variables.

analyzing ice cores can reconstruct ancient climate data with an amazing level of precision.  we can even tell pretty accurately what gases comprised our atmosphere back then and their relative proportions. tree rings also work quite well (we can also tell when ancient volcanic eruptions occurred.)

analyzing ice cores can reconstruct ancient climate data with an amazing level of precision.  we can even tell pretty accurately what gases comprised our atmosphere back then and their relative proportions. tree rings also work quite well (we can also tell when ancient volcanic eruptions occurred.)

I think you're somewhat overconfident when saying "amazing level of precision".  How do you know that?

Calibration is not straight forward.  There are no direct measurements to compare with.  There is a long mixing period (low temporal resolution).  And dating must be accurate.  Then an ice core is a sample for only one location.  And that location might not have been located precisely the same place, and the elevation will also have varied through the centuries.  I can think of a number of things that affect the accuracy.  This is not to say that ice cores don't give much useful information, but you can't compare it with direct temperature measurements, and even those are difficult to get with an "amazing level of precision".

Tree rings have their own set of problems.  They only record the growth season.  And growth depends on a number of factors: temperature, moisture, sunlight, CO2, soil, disease, and so on.  The divergence problem/"hide the decline" controversy highlight some of the issues.

We can certainly extract clear indicators for what the past climate was like based on ice cores, tree rings, pollen, sediments, and so on, but the precision is limited, and the temporal resolution is often low.

An interesting local weather phenomenon is about to strike my area -- the Fraser gap wind (also sometimes called outflow winds or squamishes). When cold arctic air comes down out of Canada, it gets dammed up on the east side of the Cascade mountains, except at narrow terrain gaps like fjords, and in particular along the Fraser River valley.  Tomorrow the air will be forced through there at up to 55 knots, which shows up dramatically in the high resolution model forecasts:

[center][/center]

In extreme cases the Fraser gap wind can sustain hurricane force, and drop temperatures from a typical 5C to below -20C in just a few hours.

By "amazing level of precision" I meant in a relative sense compared to how long ago we're talking about.

By "amazing level of precision" I meant in a relative sense compared to how long ago we're talking about.

Well, it's interesting data in the sense that climatic changes do get recorded.  But you can't really compare that with today's measured climate with much precision.  You get some ideas what whole centuries were like, but not that much better.

I heard that snow has been making it down to sea level in Washington State!