The EPA claims that North American snow cover is decreasing, when in fact winter snow cover is increasing. The past winter had the greatest snow cover on record according to Rutgers University.

February 13, 2010 – Snowcover in at least 49 states, possibly all 50.

I remember how happy I was the day the Berlin Wall fell down. Little did I realize that it was just moving west by 4,000 miles.

### Like this:

Like Loading...

*Related*

Hi Steve;

The EPA graph of NA winter snow cover is dramatically different from the graph that you show. I note that their data is in millions of square miles and the data in this article is in millions of square kilometres (if I’m reading correctly).

As an example: at a quick glance the EPA data for 1978 shows about 3.75 million square miles of snow cover.

The chart in this article shows approximately 17.8 million square kilometres of snow cover. Using 1.6 km = 1 mile then 3.75 million square miles would equal 6 million square kilometres of snow cover (I think).

So for these years the data differs by a factor of 3. If I’m `not just being stupid’, then we need a reference for the data that you give as the values dramatically differ. Is their a different methodology, etc in the data collection. I am not at all questioning what you are saying, just some additional explanation and clarity is needed.

Thanks for putting up with my questions.

Philip

Do you have a reference for the EPA graph? The graph I included is the the Rutgers University graph, which I think is the most commonly used standard.

Holy crap! When I opened the publication you linked to, went to the snow and ice page, they showed the graph and said it was from the Rutgers University graph! It’s probably well worth checking out ….perhaps they homogenized their data? Could make a good post.

I assume you would prefer 49 states to be covered in crude oil.

I live near an oil field and have not noticed anything covered with crude.

I’ll bet you don’t drive a car, or use any products dependent on fossil fuels. Like food or a computer.

Shallow Hal!!

Mystery solved. They are taking the yearly average by week of area

afterconverting km^2 to mile^2 (NA w/o Greenland).Year Miles^2 Avg Area

1972 3398466.62801896

1973 3304521.08742886

1974 3303639.84142784

1975 3395082.70000336

1976 3275434.78919124

1977 3247550.9739503

1978 3641620.51110914

1979 3364528.57006081

1980 3187625.08757446

1981 3192534.60174284

1982 3333549.79209454

1983 3295939.67738519

1984 3171074.33593349

1985 3587948.61096705

1986 3328981.14178153

1987 3042938.45179999

1988 3066195.15442613

1989 3193544.41976255

1990 3036104.60447491

1991 3168247.2959308

1992 3224172.6813195

1993 3166354.01182265

1994 3124860.93205817

1995 3221748.15281678

1996 3363519.42771987

1997 3387544.666692

1998 3027734.93928986

1999 3034537.44798976

2000 3249022.47362574

2001 3240346.951181

2002 3334801.92139373

2003 3291399.1160042

2004 3268791.32240535

2005 3115889.07620697

2006 3093593.02792573

2007 3204363.22746291

2008 3278591.72131297

2009 3262269.79955112

I just did a linear fit through that data and it is increasing , not decreasing.

Final set of parameters Asymptotic Standard Error

======================= ==========================

a = 1631.55 +/- 2167 (132.8%)

b = 69.9804 +/- 4.313e+006 (6.164e+006%)

Steven,

I believe your analysis is in error. Just looking at the data, one can see a very tiny downwards slope. I am assuming that are using a slope times (x) function which will always come out positive if (x) is positive (time series in this case).

`---------------------------------------------------------------------------------------`

[10/1/10 3:44 PM Plot: ''Graph1'']

Linear Regression of dataset: Table1_2, using function: A*x

Weighting Method: No weighting

From x = 1.972000000000000e+03 to x = 2.009000000000000e+03

A (slope) = 1.631588210518266e+03 +/- 1.177534253223456e+01

--------------------------------------------------------------------------------------

Chi^2/doF = 2.087700712620697e+10

R^2 = -0.107016561903376

Adjusted R^2 = -0.137767021956248

RMSE (Root Mean Squared Error) = 144,488.778547702

RSS (Residual Sum of Squares) = 772,449,263,669.658

---------------------------------------------------------------------------------------

With slope times (x) plus intercept, it does show a ever so slight decline in the yearly average:

`[10/1/10 3:30 PM Plot: ''Graph1'']`

Linear Regression of dataset: Table1_2, using function: A*x+B

Weighting Method: No weighting

From x = 1.972000000000000e+03 to x = 2.009000000000000e+03

B (y-intercept) = 1.159183198494020e+07 +/- 3.856522934080974e+06

A (slope) = -4.191813036342162e+03 +/- 1.937435022559441e+03

--------------------------------------------------------------------------------------

Chi^2/doF = 1.715232408531104e+10

R^2 = 0.115068551033047

Adjusted R^2 = 0.0645010396635071

RMSE (Root Mean Squared Error) = 130,966.881635439

RSS (Residual Sum of Squares) = 617,483,667,071.197

I am sure it was an auto-function in your stats package.