Dr. Ricky Rood's Climate Change Blog

Growing Season is Longer

Posted by: RickyRood, 20:36 GMT le 29 Mars 2008

Getting Ready for Spring (4):

Spring is here, and in Ann Arbor it snowed again last Thursday. It will melt fast. It’s been a winter of snow, and fog, and flood, and thunderstorms. Almost like it has been “spring” all winter.

This series of blogs (see links below) started with the revision to the planting guidelines that was published last year. This is the little map that appears on the back of the seed packages, and it has shown a shift of the planting zones. Warmer zones are creeping northward. Also in those blogs was a discussion of birds migrating and plants blooming earlier in central Europe. Then, there was a discussion of spring time snow cover, first at a single station in Alaska, then with a perspective on larger geographic regions. Large portions of the northern hemisphere have seen a reduction of springtime snow cover. There has been some regional increase in springtime snow cover, especially in the Himalaya. As some pointed out in the comments this is, in fact, consistent with increased atmospheric moisture in a warmer atmosphere and increased participation. (Note, I said consistent with, not attributable to.) In the eastern half of North America there have been patchy increases of snow cover in the spring. (Assignment: What are the climate predictions for eastern Canada and northeastern United States in March, April, May?)

Let’s return to vegetation. Satellites are able to measure vegetation in several different ways. One of the most used parameters is the Normalized Difference Vegetation Index ( NDVI). This measurement takes advantage of the fact that leaves in active photosynthesis have strongly different reflective characteristics than non-leafy environments. Plants use solar radiation in a specific part of the spectrum for photosynthesis. ( Photosynthetically Active Radiation) Outside of this range plants scatter and reflect solar radiation. Hence there is a strong edge in the spectrum of radiation reflected by the Earth. This can be used to tell when leaves come out and start to grow.

In a 2001 paper in the Journal of Geophysical Research , Liming Zhou and colleagues looked at trends of NDVI between 1981 and 1999. These data showed strong greening in the northern hemisphere between 40 degrees and 70 degrees north. Springtime moved earlier and the start of autumn moved later.

Figure 1 is from the web site of Ranga Myneni at Boston University. He is one of the co-authors of Zhou et al. paper.

Figure 1: Increase in length of growing season in North America and Eurasia. (From the web site of Ranga Myneni at Boston University. After Zhou et al. 2001)

In this paper they document that spring, north of 40 degrees in North America occurs about 8 days earlier at the end of the record than at the beginning. (The uncertainty is + or - 4 days.) In Eurasia spring is 6 (+ or – 2) days earlier. In North America fall starts 4 (+ or – 3) days later and in Eurasia it starts 11 (+ or – 3) days later. The growing season defined on hemispheric scales has increased by 12 (+ or – 5) and 18 (+ or – 4) days in North America and Eurasia, respectively. As you can see in the figure, there is variability in the time series, but the trend, which is almost half a month, is clear.

Hence, we see that throughout the northern hemisphere the beginning of leaf growth, the loss of snow cover, and the behavior of birds is consistent with the temperature increase that has been directly measured. I assert that these changes at the seasonal transition represent a natural “averaging” of the variability associated with weather. This is a robust indicator of warming.

r

Some interesting web resources:

Recent Changes in Vegetation (Arctic Report Card 2007)

Faster carbon dioxide emissions will overwhelm capacity of land and ocean to absorb carbon

A relevant lecture from my class

Blogs on spring getting earlier:

Getting Ready for Spring (1)
Getting Ready for Spring (2)
Getting Ready for Spring (3)
Jeff Masters blog on snowy winters

Updated: 23:12 GMT le 29 Mars 2008

Permalink | A A A

View Comments (154)

Snow Melting Earlier

Posted by: RickyRood, 03:12 GMT le 20 Mars 2008

Getting Ready for Spring (3):

The previous blogs started a discussion on spring coming earlier (Getting Ready for Spring (1) Getting Ready for Spring (2)). One of natural places to expect warming to be measured is at the seasonal transitions, and one measure of change would be changes in snow cover. A figure showed in the previous blog was from a single station, Barrow, Alaska. At Barrow, surface observations show that there is a greater than 10 day shift in the melt date over the 60 years of the record. The last 25 years of this record show enormous variability. There is a random aspect to this variability, and after averaging a strong trend is found. The attribution of this trend requires consideration of many other sources of information. The variability and the trends at a single station could be due to many factors.

Figure 1 shows snow trends from visible satellite observations from 1978 to 2006. This is for March-April-May.

Figure 1: Spring duration of snow cover: This map shows trends derived from visible satellite data, 1978-2006. Image by Richard Armstrong and Mary Jo Brodzik, National Snow and Ice Data Center, University of Colorado, Boulder. From the National Snow and Ice Data Center.

The blue shows areas where the duration of snow cover has decreased; the red is where the duration of snow cover has increased. In northern Alaska and the western United States the duration has decreased up to 3 days. Over much of Eurasia the snow cover has decreased. In the eastern part of North America there has been a small increase; in some of the high mountains of Asia there have been increases on the order of 3 days. Stephen Dery and Ross Brown in Geophysical Research Letters calculated snow trends for each week of the year for the Northern Hemisphere from 1972-2006. (Full paper from Stephen Dery's Web page) In both North America and Eurasia negative trends are calculated, with the trends in North America being larger. The behavior of the snow in North America and Eurasia vary together. The largest decrease of snow cover occurs in spring. There is some decrease in winter, and there is some increase of snow cover in fall. This accumulation of information across large expanses of land, and from different sources indicates systematic changes.

Phil Mote in a 2006 paper in the Journal of Climate examined the melting the western United State more fully. Mote used observations compiled at the United States Department of Agriculture. His analysis separates the changes that are correlated with indices that represent variability in the North Pacific Ocean. About half of the snow trend can be correlated with this variability, and half is related to a warming climate in the U.S. West. Further, Mote analyzed the behavior of precipitation in the west and the snow pack at relatively low altitudes in the Sierra Nevada and the Cascades – this is the transient snow zone. Aside from a general reduction in snow coverage, there is a change in how snow accumulates. Rather than a steady buildup, there is accumulation and melt, accumulation and melt. This profoundly changes stream flow and water availability.

Changes in snow are not an easy signal to extract. Snow and atmospheric warming form a complicated relationship. (Jeff Masters blog on snowy winters.) Snow observations from across the northern hemisphere paint a picture of decreasing snow cover. The decreases are largest in the spring, which is getting earlier.

r

Permalink | A A A

View Comments (109)

Creeping Onset of Spring

Posted by: RickyRood, 20:20 GMT le 08 Mars 2008

Getting Ready for Spring (2):

If the planet is slowly warming, it is difficult to make that determination from the day-to-day and even the year-to-year weather observations. There is large variability of the temperature, plus there are sometimes problems with the quality of observations. Both of these are especially true if only a single place is considered, like an airport being surrounded by a city – or your backyard. Because of this strong variability, we rely on temperature measurements collected over many years and from many places. We also use temperature data from different types of instruments, and there are continuous efforts of quality control to determine and correct observational errors. This collection of observations for many years and from many places and from many sources is a way to accumulate information and to remove the variability that is random. Random variability, by definition, averages out to zero if enough observations are accumulated.

The previous blog was on spring coming earlier (Getting Ready for Spring (1)). If we are seeing a warming at the Earth’s surface, there are many ways that this warming is manifested. One of natural places to look is at the seasonal transitions. In the interior of the United States, the transition from winter to spring to summer is a time of large changes in the weather. It is cold, then warm, then cold again. Those in the Southeast, where peaches are grown, know the tension every year as the peaches bloom and a cold front is on the way with potential fruit-killing frost. It is time of severe weather when cold air from the north and west is brought next to warm and moist air in the south and east. This is weather carrying out its role in defining the Earth’s climate, carrying heat from the equator to the pole.

The seasonal transitions are also a place where we would expect to see a natural accumulation of the impact of a warming trend. If there is a trend, then over the course of several years one would expect to see the onset of spring, perhaps defined by the last killing frost, to move earlier in the year. The transition from fall to winter, the first killing frost, would move later in the year. This is a place where the trend is “accumulated;” it is a natural place that random variability is “averaged out” and the existence of the trend is exposed.

The figure from paper by Gian-Reto Walther and many co-authors, entitled, The Ecological Responses to Recent Climate Change, in the previous blog highlighted changes in Germany of birds arriving earlier and trees getting their leaves earlier. Also in that paper is a short summary of similar observations from around the world. The report from Working Group II of the IPCC does an extensive summarization of evidence of the warm season getting longer all around the world.

Edges, like the seasonal transition, are important in the climate and in measuring climate change. Aside from the seasonal transition another place to look for changes is in mountains. Since it gets colder at higher altitudes we can see whether or not frost-free zones are moving to higher altitudes in mountains. We can look at whether or not frost-free zones are moving farther north in the northern hemisphere. That is what the map from the Arbor Day Foundation in the previous blog showed. It is worth pointing out, explicitly, that in all of these cases we are looking at frost, ice, because the formation of ice impacts plants, and some animals, strongly. It kills the peach blossoms, the fruit. Remember the ice edges in the physical climate system, sea ice, where there are large changes in albedo associated with the freezing of water and thawing of ice.

Another place that ice on the edges is important is snow cover. This year has been very snowy in the northern hemisphere. That it is snowy does not suggest that it is colder. If it gets warmer, it does not mean that we no longer see freezing temperatures in places like Michigan. If it gets warmer there is more water in the atmosphere, and when there is precipitation there will be more precipitation, and if it is below freezing, then that precipitation will be ice and snow. From a climate point of view it is more important to look at snow cover in the late winter and early spring. Is the snow melting earlier?

This figure from NOAA’s Arctic Change Web Site shows the trend in snow melt at the station in Barrow, Alaska.

Figure 1: The date of snow melt from Barrow, Alaska adapted from Stone et al. (2002)

The Barrow station is shown by Stone et al. (2002) to be representative of the North Slope of Alaska. The green line is a fit through the first part of the data record and shows only a small slope in the melt date. The red line, through the later part of the record, shows a large slope, and the black line shows the slope for the whole, approximately, 60 year record. The black line shows a greater than 10 day shift in the melt date, which is between the slopes calculated for the green and red lines.

Also on the figure is the dashed line derived from a statistical model which investigates which geophysical variables explain the snow melt most fully. Wintertime snowfall, springtime temperature and cloudiness are the most important variables. These variables are strongly linked to circulation patterns, specifically, the Aleutian Low and the Beaufort Sea Anticyclone.

Returning to the discussion of the above paragraphs, the last 25 years of this record show enormous variability. There is a random aspect to this variability, and after averaging a strong trend is found. The attribution of this trend requires consideration of many other sources of information.

r

Stone et al. (2002): Earlier spring snowmelt in northern Alaska as an indicator of climate change, J. Geophys. Res., 107, 10.1029/2000JD000286.

Updated: 03:29 GMT le 05 Mars 2009

Permalink | A A A

View Comments (131)

About RickyRood