Archives

Is It Cloudier Than It Used to Be?

I have had the general impression that in recent years, it has gotten cloudier. But that could be an incorrect impression, coming from the fact that, like everybody, I get older with each year. Or, it could come from the fact that my analyses for this blog have shown that there is marginally more precipitation in Missouri due to climate change. Perhaps I say “Ha!” to myself every time it is cloudy, “See, it’s because of climate change.”

Checking out one’s subjective impressions about clouds is not so easy. Clouds are very complex, and they provide the largest source of uncertainty in future climate projections. There are all kinds of clouds, and they exist at many levels of the atmosphere. Not only that, they are constantly changing: it can be overcast one moment and clear an hour later. The task of reducing this complexity to a simple measure of cloud cover has been difficult. Most weather stations do it, however. For instance, the daily climate report published by National Weather Service in St. Louis reports that for St. Louis, on July 21, 2019, the sky cover was 0.6. This means that on average for the whole day, 6/10ths (60%) of the sky was covered with clouds. (They measure the percent of the sky covered by clouds several times a day, and average the results.)

Finding a time series reporting this data over time is harder. I’ve been looking for it for some time, and I finally found it at NASA’s Giovanni Data Tool. This is a portal that provides access to satellite data for a large number of atmospheric variables, including cloud fraction. Without getting into the weeds, cloud fraction is the fraction of an area that is cloudy, roughly the same thing as sky cover, but from the sky down, rather than the ground up. The scale runs from 0, meaning none of the area was covered by clouds, to 1, which means all of the area was covered by clouds. Thus, as in the example above, 0.6 would mean that 6/10ths of the area (60%) was covered by clouds.

Figure 1: Cloud Fraction Search Area. Map Created on Google Earth.

Giovanni did not permit searching for cloud fraction by state, but it would search within a rectangle that I could define. So, as I usually do in such cases, I defined a rectangle that just barely enclosed the whole state. Figure 1 shows the search area.

.

.

.

.

.

.

.

.

 

Figure 2. Data source: NASA Giovanni.

Figure 2 shows the monthly cloud fraction for that rectangle from 1/1/1980, through 6/1/2019. To justify my impression that it is cloudier, you can notice that from about 2010, there has been a noticeable increase in cloud fraction. However, it is an anomaly, and the trend line, in black, shows that the cloudiness has not changed much since 1980; if anything, it has decreased a little bit, the trend line being down about 0.02 over the whole time period.

Thus, my impression was both right and wrong. I was right that, since 2010, it has gotten more cloudy over Missouri. But that hides the long-term fact that, since 1980, it has not. It is common for subjective impressions to favor recent experience over the far past, but it can hide the truth.

.

.

.

Figure 3. Data source: NASA Giovanni.

What about the Continental United States (CONUS) as a whole? Again, Giovanni wouldn’t let me search by the boundaries of the CONUS, so I searched in a rectangle that barely enclosed it. Figure 3 shows the data. Notice first that the year-to-year variation is less than for Missouri. Where cloud fraction in Missouri has bounced around between 0.2 and 0.6, cloud fraction for the CONUS has bounced around between 0.3 and 0.5. We have seen this before with other environmental data: large areas tend to average out variations in one area with opposite variations in another. However, here, too, we see the slightly declining trend in cloud fraction. The CONUS as a whole is becoming slightly less cloudy, and the difference in the trend line is about 0.02, similar to what it was for Missouri.

.

.

.

.

Figure 4. Data source: NASA Giovanni.

The world, however, shows a markedly different trend (Figure 4). Again, because it is an even larger area, the year-to-year variation is now only between 0.44 and 0.5. Don’t be fooled because Excel has spread out the y-axis. However, this time the trend is slightly up, the trend line increasing slightly less than 0.02. The world is becoming very slightly more cloudy. I don’t know for sure, but I would bet you that the increase is happening primarily over the oceans. That would be an interesting research project.

The changes are very small. However, the world has become slightly cloudier over the last 40 years, while both Missouri and the Continental United States have become slightly less cloudy.

Sources:

“Analyses and visualizations used in this article were produced with the Giovanni online data system, developed and maintained by the NASA GES DISC.” Data downloaded 7/22/2019 from https://giovanni.gsfc.nasa.gov/giovanni.

Map created with Google Earth on 7/22/2019.

Hurricane Barry, Tropical Cyclones, and Climate Change


Tropical Storm Barry formed in the Gulf of Mexico on July 11. It strengthened, and churned ashore as a Category 1 hurricane in western Louisiana on July 13. It weakened, and moved northward, causing rain in Arkansas and here in Missouri.


Figure 1. Data source: Landsea, downloaded 2019.

As Figure 1 shows, land-falling hurricanes in July are uncommon, though not unknown: there have been 26 in the 167 years that records have been kept. That’s one every 6.4 years. It seems like a good time to ask whether climate change has been affecting tropical cyclones?

.

.

.

.

.

.

Figure 2. Source: Weather.gov.

Why might we expect climate change to affect tropical cyclones? To answer that question, you have to understand the “engine” that drives a tropical cyclone (see Figure 2). Tropical cyclones get their energy from warm, humid air on the surface of the ocean. Convection causes the warm, humid air to rise, and as it does so, it enters cooler regions of the atmosphere. This causes the humid air to condense into clouds and rain (often thunderstorms). Condensation is an exothermic process – that means that the water gives off heat as it condenses. The heat keeps the humid air rising, condensing more water, and giving off more heat. This process continues. As the air rises, it leaves an empty place where it used to be, so more air rushes in from the sides to take its place. If this air is also warm and humid, then it will rise, too, and condense into rain. If this process strengthens, then the air rises faster and faster, and the air moving in to take its place moves faster and faster. The air begins to rotate, and presto, you have your tropical cyclone.

The energy that drives all of this is the warm humid air on the surface of the ocean. Thus, it is easy to understand that anything that causes the air on the surface of the ocean to be warmer and more humid can provide more energy to a storm that might form.

What if, over the decades, the water in the oceans got warmer? Well, it would make the air above it warmer. It would also evaporate into the air more effectively, as we all know that warm water evaporates more quickly than does cold water. So warming oceans would seem to be a perfect recipe for providing more energy to tropical cyclones, making them more intense. Climate change is projected to cause the oceans to warm, and this brings us to the first article I wanted to report.

Multiple studies have reported that the heat content of the oceans has been rising. The IPCC 5th Assessment Report put the rate at 0.20-0.32 watts per square meter. However, there were many uncertainties. A recent article by Cheng, Abraham, Hausfather, and Trenberth (2019) reports that since the 5th Assessment Report, scientists have made progress in identifying and resolving the uncertainties. They review 3 studies that incorporated the advances, and find that the rate has actually been 0.36-0.39 watts per square meter. Compared to the IPCC estimate, that represents an increase of somewhere between 0.04-0.19 watts per meter.

Doesn’t sound like much, does it? The oceans are huge, however, 361,900,000 square kilometers, which translates to 361,900,000,000,000 square meters. So, the increase represents an increase of 14,476,000 – 68,761,000 megawatts. The Callaway Nuclear Generating Station in Missouri is rated at 1,190 megawatts, so the increase is equal to 12,164 – 57,782 Callaway Nuclear Generating Stations. That’s a lot of heat!

So, have tropical cyclones become more severe? Well, that is really two questions. One involves wind speed, the other involves rainfall amounts. Too many other factors affect wind speed and rainfall amounts to permit a simple comparison across storms. There is no scientific consensus regarding how climate change has affected tropical cyclones, or how it may do so in the future.

Patricola and Wehner (2019) recently published a study where they modeled the wind speed and rainfall in a suite of 15 tropical cyclones from around the world under different climates. Thus, this study doesn’t really prove anything. Rather, it clarifies what kind of effects our current theories might predict. From coolest to warmest they simulated pre-industrial climate, historical climate, RCP 4.5, RCP 6.0, and RCP 8.5. (The RCPs are standardized emission scenarios used to project the effects of climate change. The terms 4.5, 6.0, and 8.5 represent the level of radiative forcing caused by climate change. All are projected to be warmer than current climate.) They then made comparisons between the models.

Table 1. Source: Patricola and Wehner, 2018.

Table 1 presents the results for peak wind speed measured for at least 10 minutes. The 1st column lists the name of the storm. The 2nd column gives the difference between the result of the historical and pre-industrial models. The 3rd column gives the difference between the result of the RCP 4.5 and historical models. The 4th column gives the difference between the result of the RCP 6.0 and historical models. The 5th column gives the difference between the result of the RCP 8.5 and historical models. The 6th column gives the wind speed projected by the historical model. The 7th column gives the wind speed as it was actually observed in the real storm.

Remember that the goal here is not to actually predict wind speed, but to understand the kind of effects our climate models project. The average difference in wind speed projected for RCP 4.5 vs. historical climate was 6.7 knots. The average difference in wind speed projected for RCP 6.0 vs. historical climate was 7.8 knots. The average difference in wind speed projected for RCP 8.5 vs. historical climate was 13.0 knots. Thus, for those comparisons, the hotter the climate scenario, the higher the wind speed. The outlier was for pre-industrial climate. Being cooler, the pre-industrial climate scenario should have resulted in lower wind speeds than the historical climate, yet the projection resulted in higher wind speed. Why, I’m not sure.

Table 2. Source: Patricola and Wehner, 2018.

Table 2 presents the results for rainfall. The average difference in rainfall projected for RCP 4.5 vs. historical climate was 10.9 inches. The average difference in rainfall projected for RCP 6.0 vs. historical climate was 13.5 inches. The average difference in rainfall projected for RCP 8.5 vs. historical climate was 18.4 inches. Here, pre-industrial climate was not an outlier: the average rainfall projected for pre-industrial climate was 5.8 inches less than for historical climate.

Thus, the study shows that modeling projects that climate change will result, on average, in more rainfall per storm. The trend was linear and consistent, and no storm bucked the trend. For wind speed, results were less consistent.

Hurricane Harvey was a great example of what is projected for rain. The storm parked itself over Houston and Eastern Texas, dropping 40 inches of rain in some areas, causing extensive flooding.

But what about Hurricane Barry? It’s storm path initially projected that it would come very close to New Orleans, and it would dump 15-20 inches of rainfall. Would there be another Katrina-like disaster?

The reality turned out to be much less dire. The storm came ashore west of New Orleans, and though there were some spots that received heavy rain, most areas received much less. According to the official weather service reports from New Orleans, Baton Rouge, and Shreveport, between July 1 and July 15 they received 3.66 inches, 4.41 inches, and 0.42(!) inches. It sounds like most of the rain came from scattered thunderstorms associated with Barry, not from a widespread downpour.

Sources:

Cheng, Lijing, John Abraham, Zeke Hausfather, and Keven E. Trenberth. 2019. “How fast are the ocearns warming?” Science. Vol 363 (6423), pp. 128-129. Downloaded 1/20/2019 from http://science.sciencemagazine.org.

Landsea, Chris. “Frequently Asked Questions: How many hurricanes have there been in each month?” Atlantic Oceanographic & Meteorological Laboratory. Data downloaded 7/16/2019 from https://www.aoml.noaa.gov/hrd/tcfaq/E17.html.

National Centers for Environmental Information. “Volume of the World’s Oceans from ETOPO1.” Viewed online 7/15/2019 at https://ngdc.noaa.gov/mgg/global/etopo1_ocean_volumes.html.

National Weather Service Forecast Office, New Orleans/Baton Rouge, LA. Daily Climate Report.Viewed online 7/16/2019 at https://w2.weather.gov/climate/index.php?wfo=lix.

National Weather Service Forecast Office, Shreveport, LA. Daily Climate Report. Viewed online 7/16/2019 at https://w2.weather.gov/climate/index.php?wfo=shv.

Patricola, Christina M. and Michael F. Wehner. 2018. “Anthropogenic Influences on Major Tropical Cyclone Events.” Nature. 563, 11/15/18., pp. 339-346.

Weather.gov. “Hurricane Facts”. Downloaded 7/16/2019 from https://www.weather.gov/source/zhu/ZHU_Training_Page/tropical_stuff/hurricane_anatomy/hurricane_anatomy.html.

Wikipedia contributors, “Hurricane Barry (2019),” Wikipedia, The Free Encyclopedia, https://en.wikipedia.org/w/index.php?title=Hurricane_Barry_(2019)&oldid=906405193 (accessed July 15, 2019).

Wikipedia contributors, “Tropical cyclone,” Wikipedia, The Free Encyclopedia, https://en.wikipedia.org/w/index.php?title=Tropical_cyclone&oldid=906385211 (accessed July 15, 2019).

Western USA Snowpack and Reservoirs


I last reported on the snowpack and reservoirs in California and the Colorado River Basin at the end of January. At that time, the snowpack had gotten off to a slow start. Boy, did things change!


Most of the western USA has a monsoonal precipitation pattern: most of the rain falls during the winter, and from about March through the end of November, there usually isn’t much precipitation. For that reason, the entire region depends on stored water, either in reservoirs, or perhaps even more importantly, in the snowpack that builds up on the mountains. Because I have family in California, and because the long-term predictions for the water supply in the West have been grim, I have been following the status of the snowpack and the reservoirs in California and the Colorado River Basin. The most important measurement for the snowpack is usually around April 1, while maximum levels on the reservoirs typically occur some weeks later, as the snowpack melts. For instance, Lake Powell, the uppermost very large reservoir on the Colorado River, reaches its highest monthly average in July. Lake Shasta, the largest reservoir in California, reaches its highest average level in May.

Figure 1. Source: California Department of Water Resources, 2019a.

Figure 1 shows the snow water content of the snowpack in California. (The snow water content represents how much water there would be if you melted the snow in a given location. For instance, if you melted 7 inches of snow, it might only represent 1 inch of water.) The 3 charts represent the 3 major snowpack regions of California. The dark blue line is for 2019, while the light blue area represents average. The units along the y-axis represent percent of the April 1 average.

You can see that 2019 had an above average snowpack, maxing out at more than 150% of average in all 3 regions. By this time of year, the snowpack has largely melted. Notice the text at the bottom right: “Statewide Percent of Average for Date: 71%.” Despite having a snowpack that maxed out at 150% of average, the amount of snowpack remaining on this date is less than average. This illustrates another way that climate change is affecting California: temperatures are up, and the snowpack is melting more rapidly than in the past.

Figure 2. Source: Mammoth Mountain Ski Area, 2019.

I use Mammoth Mountain to illustrate snowfall amounts; it is located in the middle-south of the Sierra Nevada Mountains, south of Yosemite. Their website indicates that they are still open with 15” at the main lodge and 55” at the summit – on July 4th! Figure 2 shows snowfall at Mammoth Mountain by year and month. Paralleling the snowpack survey, this chart shows that 2019 was well above average at Mammoth, but not a record. There was one month with a lot of snow: February.

.

.

.

Figure 3. Source: California Department of Water Resources, 2019b.

As a result, California’s reservoirs are all above average for this date, as shown in Figure 3. Even lake Oroville, which had to be mostly emptied when the dam eroded, threatening catastrophic failure, is nearing capacity. Every single reservoir is above average, and only one is not near capacity.

Many western states, including Southern California, are heavily dependent on water from the Colorado River. Lake Mead is the largest reservoir, and it is capable of holding more water than any other in the USA. For a couple of decades, there has been concern that water demands on the Colorado River had increased, and water supply into it had decreased, to the point that Lake Mead would be drained within a couple of decades. Over the last 5 years, water levels were so low that they flirted with the mandatory cut-back level: states would have lost a significant portion of their water.

.

.

.

Figure 4 shows that snowpack in the Upper Colorado River Basin was higher than average in 2019, and this represents the 2nd time in the last 3 years. This snow melts into a number of reservoirs along the tributaries of the Colorado River, and then into Lake Powell, the first of the gigantic reservoirs along the Lower Colorado River Basin. From Lake Powell, it is released into Lake Mead. Figure 5 shows that Lake Mead is up from its record low a few years ago, but it is still historically very low.

Figure 4. Source: water-data.com, 2019b.

Figure 5. Source: water-data.com, 2019a.

 

 

 

 

 

 

 

 

 

 

 

The bottom line here is that the draught has finally broken in California, and that state is sitting on plenty of water for now. This was to be expected – nobody ever thought that California’s water supply problem would be a straight line from full to empty. The regions history, however, indicates that draught is a normal occurance for the state, and in recent years, wet periods have not lasted too long. The whole point of reservoirs is that they get drawn down during dry periods. So long as they get refilled before they are empty, the system is working just like it should. Three things can break the system, though: first, if water demand increases too much, and there just isn’t enough water to satisfy the demand. California is getting close to outstripping supply, as is all of the West. Second, wet years could get less wet, and then they might not be sufficient to refill the reservoirs. This year was sufficient to refill the California reservoirs, but Lake Mead still has a long way to go! And finally, if too many years go by before a wet one comes along, then the reservoirs could get sucked dry. California was getting close, and Santa Barbara, in particular, got really, really close.

The long term projection is still guarded, as population continues to enter western states and climate change continues to threaten the snowpack. How it will unfold year-by-year is anybody’s guess, but for now, things are better than they were a couple of years ago.

Sources:

California Department of Water Resources. 2019a. Current Reservoir Conditions. Downloaded 7/4/2019 from http://cdec.water.ca.gov/reportapp/javareports?name=rescond.pdf.

California Department of Water Resources. 2019b. California Snow Water Content, July 1, 2019, Percent of April Average. Downloaded 7/4/2019 from http://cdec.water.ca.gov/reportapp/javareports?name=PLOT_SWC.pdf.

water-data.com. 2019. Lake Mead Water Levels, All Time. Downloaded 7/4/2019 from http://graphs.water-data.com/lakemead.

water-data.com. 2019. Lake Mead Water Level: Averages by Month. Downloaded 7/4/2019 from http://lakepowell.water-data.com/index2.php.

water-data.com. 2019. Upper Colorado Basin Snowpack (Actual Values). Downloaded 7/4/2019 from http://graphs.water-data.com/ucsnowpack.

Mammoth Mountain Ski Area. 2019. Extended Snow Report. Downloaded 7/4/2019 from https://www.mammothmountain.com/winter/mountain-information/mountain-information.

Land Use Change

This week I’m sharing with you some land use-land cover (LULC) maps. LULC maps show what is on a given area of land: water, developed cities, forests, pastures, crops, etc. I’ve been looking for recent LULC maps of Missouri for some time, and only just now found them at the website of the Multi-Resolution Land Characteristics Consortium (MRLC). This is a cooperative project by several branches of the federal government. Modern LULC maps are created by sophisticated imaging systems on satellites. This process allows for new maps to be created every few years, and the source I found has maps for 2001, 2003, 2006, 2008, 2011, 2013, and 2016.

LULC data is used for scientific study and for planning. For the purposes of this blog, what would be most interesting would be data regarding how much land in Missouri has changed use, and whether the changes represent a degradation of environmental carrying capacity. It appears to me that such data is available, either at MRLC or at the U.S. Geological Survey, but that it requires special programs and knowledge to utilize, which I don’t have. The closest I can come is a map provided by MRLC that shows the Land Use Change Index (land that changed use at least once between 2001, 2003, 2006, 2008, 2011, 2013, and 2016).

Figure 1. Missouri Land Use-Land Cover. Source: MRLC.

Figure 1 shows the 2016 LULC map for Missouri and a little bit of neighboring states. Each pixel on the map is 30 x 30 meters, which is close to 100 by 100 feet. Each pixel is coded with a color representing 1 of 21 categories of land use, shown in the legend. The pinks and reds jump out visually; these represent developed land. It is easy to make out the St. Louis, Kansas City, Springfield, Joplin, St. Joseph, Columbia, and Jefferson City areas, as well as other, smaller developed areas. The brown areas represent cultivated crops. The Bootheel is very heavily cultivated, but so is much of northern Missouri, especially the northwestern part. Green stands for forests – dark green for conifer forest and light green for mixed forest. There is a lot of forest in the Ozarks, mostly mixed forest. Yellow stands for pastureland/hay, which is widely distributed throughout the state. Blue stands for water bodies.

Looking at the map and forming overall impressions, I was first struck by how fragmented land use is. Land use varies by small plots in many areas of the state. Second, I was struck by how the forest land in the state is concentrated in the Ozarks. There are small forested regions in the north, but no very large tracts like in the south. Similarly, I was struck by how cultivated crops are almost completely absent from the Ozarks. There may be some farming there, but it looks to be devoted to hay and pasture.

Figure 2. Missouri Land Use Change Index. Source: MRLC.

Figure 2 shows a map of the Land Use Change Index in the same area. The resolution of this map is the same, each pixel is 30 x 30 meters. The map codes each pixel according to whether the use changed at least once between the mappings in 2001, 2003, 2006, 2008, 2011, 2013, and 2016. The change could have occurred in either direction. For instance, blue codes water change: what was a body of water became something else, or something else became a body of water. There are orange dots across the top of Missouri: either land became hay/pasture, or hay/pasture became something else. In addition, the change doesn’t have to have been permanent; what was hay/pasture in 2003 could have become cultivated cropland in 2006, but then reverted back in 2008. The map doesn’t tell us, it only tells us that at some point the land changed.

It’s not clear to me why they would construct this index this way. Perhaps these changes are reversible only in theory. For instance, the pink dots represent land that has either either urbanized or “un-urbanized.” I think that in real life, it is nearly impossible for land to “un-urbanize” – streets and buildings don’t go away in a matter of a couple of years. But it would be much more useful to know which changes were permanent, and which represented a degradation of the land.

Dark green represents land that did not change use across the 12-year time period. Most of the state is dark green. However, it is surprising how much land has changed. There are a lot of orange dots across northern and western Missouri, indicating that a lot of land there either became or stopped being hay fields/pastureland. Across southern Missouri, there are a lot of light green dots, indicating lots of change in the forests there.

Pink dots are scattered around the St. Louis, Kansas City, and Columbia regions, indicating land that (probably) urbanized. Given all of the development that has occurred in Springfield and Branson, I’m surprised that there isn’t more pink in that region.

Figure 3 shows the Land Use Change Index for the eastern Continental United States (CONUS). What immediately stands out is the forestland change across the Southeast, the Far North, and Maine. The map shows that change is happening, but not what kind of change. Figure 3 also suggests that, while land use is changing more in Missouri than in some states, it is changing less than in others.

Figured 3. Land Use Change Index, Eastern CONUS. Source: MRLC.

Figure 4. Land Use Change Index, Western CONUS. Source: MRLC.

 

 

 

 

 

 

 

 

 

 

Figure 4 shows the Land Use Change Index for the western CONUS. Across the West, forest change seems to be the most common change, followed by persistent grass and shrub change. One wonders how much of this is due to logging and/or development. However, one also wonders how much of this is due to either the bark beetle infestations that have been killing trees all over the West, or to the huge wildfires that have been ravaging these areas? For instance, using the interactive map on the MLRC website, I can zoom in and identify forest changes inside Glacier National Park. Those are not logging changes, they have to be the result of wildfire. However, there have not been hundreds of wildfires all up and down the Cascade Mountains in Oregon, those are likely to be logging changes.

Maps like these are created using large databases; each pixel is identified and has data coded for it. Theoretically, one should be able to use the database to construct summary statistics. I found a couple of databases, but unfortunately, they required specialized software to use, and/or they were multiple terabytes in size. Further, the agencies that create these maps don’t seem to think that states make useful boundaries. I will continue to keep my eyes open, however, and if I should come across summary data, I will do a future post on it.

Sources:

Land cover change 2001-2004-2006-2008-2011-2016. Multi-Resolution Land Characteristics Consortium. Downloaded 2019-07-02 from https://www.mrlc.gov/viewer.

Screen capture from Multi-Resolution Land Characteristics Consortium. 2016 CONUS Land Cover. Downloaded 2019-07-02 from https://www.mrlc.gov/viewer.

The World’s Thinning Glaciers


Glaciers around the world are melting. Millions of people around the world who depend on them are likely to be impacted.


One of the signs of climate change that has received the most attention is the shrinking of glaciers around the world. Sometimes it is presented as a cause of sea level change, but it has only a minor effect on sea level. The Greenland Ice Cap and the Antarctic Ice Cap are far larger bodies of ice, and they will (and already do) contribute more to rising sea levels than do all the glaciers around the world. Further, much of the predicted rise in sea level is due to nothing more than the thermal expansion of water. You know, things expand as they heat up. Well, the oceans are projected to heat up only a little, but there is so much of them that expansion contributes significantly to the rise in sea level.

Melting glaciers matter for a different reason: people depend on them for water. Glaciers form the headwaters of many of the world’s rivers, great and small. Not meaning to make a comprehensive list, in Asia, the Indus, the Ganges, the Brahmaputra, the Yangtze, the Huang-ho (Yellow), and the Oxus all arise from glacial melt. In Europe, the Danube, the Rhine, and the Po all receive substantial glacial melt. In South America, the Madeira (largest tributary of the Amazon) receives glacial melt from about 1,000 miles of the east slope of the Andes. Finally, in North America, the Missouri, Columbia, Snake, Yukon, McKenzie, and Fraser Rivers all receive significant glacial melt.

Figure 1. Source: Schaner, Voisin, Nijssen, and Lettenmaier, 2012.

Figure 1 is a map indicating river basins for which at least 5% (green), 10% (yellow) 25% (orange), and 50% (red) of discharge is derived from glaciers in at least one month. (The “at least one month” qualification matters – glaciers melt much more during the warmer months of the year). Notice that one of the 2 largest blotches of color is located along the northwest coast of North America. This is a high mountain region that is very far north and close to an ocean: a perfect recipe for glaciers. The other is located in Central Asia, where the highest mountains in the world are located, and which receive the famous monsoons of India.

.

.

.

.

.

Table 1. Source: Schaner, Voisin, Nijssen and Lettenmaier, 2012.

Table 1 shows the number of people and the land area that depend on glacial melt. Considering the world as a whole, an estimated 120 million people depend on rivers that get 50% or more of their water from glacial melt (1.8% of the world’s population). About 600 million people depend on rivers that get 5% or more of their water from glaciers (8.9%of the world’s population). So, we are talking about substantial numbers of people. Should the earth’s glaciers decline substantially, some of these people would be likely to lose access to water entirely, at least for part of the year. For others, important life-sustaining activities, such as agriculture or transportation, would be curtailed.

.

.

.

Figure 2. Source: WGMS, 2017, updated, and earlier reports.

So what is the status of the world’s glaciers? Sadly, it is not good! The World Glacier Monitoring Service (WGMS) is a joint project of the World Data System, the International Association of Cryospheric Sciences, the United Nations Environment Program, the United Nations Education, Scientific, and Cultural Organization, and the World Meteorological Organization. The WGMS studies and monitors the world’s glaciers, and serves as a repository for data on them. They have a set of 30 glaciers around the world that have been repeatedly measured for at least the last 30 years (some much longer), with few or no gaps. Figure 2 shows the status of these 30 glaciers. The year is represented on the x-axis, and the change in mass is represented on the y-axis. The units on the y-axis are meter water equivalents, which are equal to metric tons per square meter of surface. Thus, in 2015, the year of greatest loss, these 30 glaciers collectively lost about 1.1 metric tons of ice per square meter of surface. When you consider that the earth has hundreds, if not thousands, of glaciers, then it becomes clear that we are talking about a lot of ice that is melting into water.

.

Figure 3. Source: WGMS, 2017, updated, and earlier reports.

Many of these glaciers are hundreds or thousands of feet thick, and the loss in mass represents thinning of the glacier (melting from the top or bottom) every bit as much as it represents retreat (melting at the bottom end of the glacier). Figure 3 shows the cumulative loss in mass of these same 30 glaciers since 1950. Don’t be confused by the early values above 0 – the glaciers have been losing mass throughout, but for some reason, the WGMS set 1976 as zero, not 1950.

.

.

.

.

Figure 4. Source: WGMS, 2017, updated, and earlier reports.

The reference glaciers are concentrated in North America and Europe more than in other continents. However, consider Figure 4, which shows the cumulative mass lost by region. Western Canada/USA and Central Europe have had greater loss than any other regions. However, all regions have had significant loss, including Svalbard and Jan Mayen (3rd worst), and Asia Central (4th worst).

I thought I would illustrate the global nature of the retreat with reference to a few very well known glaciers. Though not necessarily the largest or most important, they are famous.

.

.

.

Figure 5. Mt. Everest and the Khumbu Glacier. Source: NASA 2011.

To represent Asia, I chose the Khumbu Glacier. Located in Nepal, this is the glacier of Mt. Everest. Base camp sits on it; climbers walk up it and through the Khumbu Ice Fall (where the glacier pours over a cliff), before starting their ascent of the mountain itself. It was measured 3 times: 1970, 2000, and 2016. Between 1970 and 2000, it thinned by an average of 300 cm. (9.8 feet) per year. Between 2000 and 2016, it thinned faster, by an average of 500 cm. (16.4 ft.) per year. (The surface of most glaciers collect dust and debris, thus parts of the glaciers turn brown or gray.)

.

.

.

.

.

.

Figure 6 Photo by John May, 2015.

To represent Europe, I chose the Mer de Glace, the famous glacier just east of Mt. Blanc (and the 2nd largest in Europe). The first measurement of the Mer de Glace was in 1570. I told you some of these measurements went back more than 30 years! By the early 1600s, the front of the glacier had advanced by about 1,000 meters. It then varied until the late 1800s, when it began retreating. By the early 2000s, the front of the glacier had retreated about 1,000 meters from its 1570 location, and about 2,000 meters from its location during the mid-1800s. Meanwhile, the thickness of the glacier was measured in 1980, 2003, and 2012. Between 1980 and 2003, it thinned at a rate of about 18-20 mm. per year (0.06-0.065 ft.) Between 2003 and 2012, the thinning accelerated to about 160 mm. per year (0.5 ft.).

Figure 7. Source: NASA.

To represent North America, I chose the Muir Glacier: the photos of its retreat are as dramatic as any around the world. It was first measured in 1880, and since then its front has retreated about 29,000 meters (95,144 ft. or 18 miles). The photos in Figure 7 were taken in 1941 and 2004, and show about 7 of those 18 miles of retreat.

I’ve discussed what climate change and snowpack loss in the Northern Rockies might mean for the water supply in the Missouri River, and those who want to explore that topic can find the post here.

Glacial loss matters in some locations more than others. A very large number of people are likely to be affected, especially in Asia. Those people often live at a subsistence level; what loss of the glaciers will mean to them is hard to know. What kind of famine, pestilence, migration, political instability, and war might result is anybody’s guess.

Sources:

NASA. 2011. Adapted from ”everest_ali_2011298_geo.tif.” Downloaded 2019-07-01 from https://visibleearth.nasa.gov/view.php?id=82578.

NASA. “Graphic: Dramatic Glacier Melt.” Global Climate Change. Downloaded 6/24/2019 from https://climate.nasa.gov/climate_resources/4/graphic-dramatic-glacier-melt.

Schaner, Neil, Nathalie Voisin, Bart Nijssen, and Dennis P. Lettenmaier. 2012. “The Contribution of Glacier Melt to Streamflow.” Environmental Research Letters. 7 034029. Downloaded 6/24/2019 from https://iopscience.iop.org/article/10.1088/1748-9326/7/3/034029.

WGMS. 2019. WGMS Flucuations of Glaciers Browser. Data accessed online 6/24/2019 at https://www.wgms.ch/fogbrowser.

WGMS (2017, updated, and earlier reports): Global Glacier Change Bulletin No. 2 (2014-2015). Zemp, M., Nussbaumer, S. U., Gärtner-Roer, I., Huber, J., Machguth, H., Paul, F., and Hoelzle, M. (eds.), ICSU(WDS)/IUGG(IACS)/UNEP/UNESCO/WMO, World Glacier Monitoring Service, Zurich, Switzer- land, 244 pp., based on database version: doi:10.5904/wgms-fog-2018-11. Downloaded 6/24/2019 from https://wgms.ch/global-glacier-state. (While this is the citation the source document suggests, the graphs used in this post were updated in January, 2019.)

Peak Streamflow Increasing in Missouri

Missouri and other parts of the Midwest are experiencing severe flooding, perhaps historic flooding. The record flood in this part of the country occurred in 1993. According to Chris Boerm, transportation manager for Archer Daniels Midland, the 1993 flood was concentrated in Iowa and the upper Midwest. This one is more expansive, affecting the entire Mississippi River, the Arkansas River, the Illinois River, the Ohio River, and the Missouri River. (quoted in Sullivan, Singh, and Bloomberg, 2019). Some 203 river gages along U.S. rivers are at or above flood stage.

With every flood, it seems, we hear a chorus complaining that flooding is getting more severe, and that our efforts to manage our major rivers have actually made things worse. Flood plains upstream act as sponges, absorbing flood water and then releasing it slowly over time, thus reducing the severity of flooding downstream. But levees along the river prevent this from happening, funneling all of the water downstream, worsening flooding there.

I thought I would look and see if, indeed, there is a trend towards increased flooding, and if so, how severe it was.

First I decided that I would focus only on Missouri. Then I decided that I would focus only on select rivers that represented diverse geographical areas of the state. Then I decided I would focus only on rivers that were relatively major rivers. And finally, I decided that I would eliminate rivers that I felt were almost entirely controlled by dams. The White River, for instance, is one our longer rivers, though it only flows through Missouri for part of its length. While in Missouri, it is impounded by 3 reservoirs: Table Rock Lake, Lake Tanneycomo, and Bull Shoals Lake. So, I eliminated it, and other similar rivers. I did not, however, eliminate the Missouri and the Mississippi. Though those rivers are regulated by dams and impounded into reservoirs, their many floods indicate that they are not almost entirely controlled by anything.

Figure 1. Location of USGS river gages. Source: USGS Mapper.

But how to measure flooding? I decided to use two measurements routinely made by the United States Geological Survey at thousands of river gages, which cover every major river in the country: peak streamflow, and peak gage height. Peak streamflow is the highest amount of water flowing down the river at any given time during a water year (water years begin in the summer). Peak gage height represents the highest the river is during a water year. These two measurements are not specific indicators of flooding. However, high readings go along with flooding, and if these two measurements are increasing, it would provide support for the idea that floods are getting worse.

Figure 1 shows a map of the river gages I selected for my study. They included gages on the Mississippi River at Grafton and at Thebes, a gage on the Missouri River at Kansas City, gages on the Meramec River near Steelville and near Eureka, a gage on the Gasconade River at Jerome, a gage on the Grand River at Sumner, a gage on the Pomme de Terre River at Polk, and a gage on the Current River at Van Buren.

Each gage has historical data for peak streamflow and peak gage height for each water year. How far back the data goes varies between gages. I turned this data into graphs, shown as Figures 2-10. For each graph, streamflow is shown in orange, and should be read against the left vertical axis. Gage hight is shown in blue, and should be read against the right vertical axis. I had Excel drop linear regressions on each of the lines, to show the trend over time. They are shown as dotted lines. I will discuss the results after sharing the charts.

(To view a chart, click on it. Once a chart is open, you may cycle through the charts by using the buttons below the charts. To return to this post from the charts, click on the name of the post under the chart.)

As one considers the charts as a group, the most obvious thing that jumps out is the large variation in streamflow from year-to-year. This is particularly evident on smaller streams that don’t gather precipitation from large drainage areas. The Grand River, for instance, had a minimum streamflow of 6,320 cfs in 2003, but a maximum streamflow of 180,000 cfs in 1947. The maximum streamflow was more than 28 times the minimum. However, even on the big rivers the yearly variation was large: on the Mississippi River at Thebes, the minimum was 140,000 cfs in 1934, while the maximum was 1,050,000 in 2016 (7.6 times the minimum).

There are 18 trend lines: 2 lines for each of 9 gage locations. All but 1 show an increasing trend over time. The only trend that isn’t upward is streamflow on the Meramec River near Steelville. I’m not sure what this means, as the gage height there does trend up, and both streamflow and gage height on the Meramec near Eureka also trend up. Eureka is downriver from Steelville. This one finding notwithstanding, with 17 out of 18 trending upward, I think it is safe to say that both streamflow and gage height have been increasing over time in Missouri.

Don’t read too much into the steepness of the different trendlines, they are determined by the scales Excel chose for the vertical axes.

At each location peak streamflow and peak gage height tend to vary within a limited range, but this range is broken in some years by extremes. Even high values in the normal range may go along with flooding in some locations, but the extremes probably indicate more severe flooding. If there is an upward trend in the normal range, it may indicate a trend toward increased minor flooding. But if there is an increase in the extremes, it may indicate that extreme flooding is getting even more extreme. And that is what we find. On most of the charts, the extreme peaks on the right are taller than the extremes on the left.

Put this together with increased development in flood plains, and yikes! The levees better hold!

The trend is not universal, however, and one of the locations that turned out to be more complex was the Missouri River at Kansas City. The highest streamflow there occurred in 1951, and streamflows since then (even in 1993) were lower. Gage height, however, peaked in 1993. The series of dams on the Missouri River were completed in 1962, and they may have moderated streamflow since then. (Although when flooding is extreme, the dams have to dump water to prevent themselves from being overtopped, and that can make things worse. See my posts on Oroville Dam.)

(Added note 6/27/19: This may actually be an effect of levee building. Levees constrict the width of the river during high water. If the river width is sufficiently narrowed, the gage level might be considerably higher, but the river might still be carrying less water.)

So, it was a lot of work to find this data and put these charts together. But they do tend to support the notion that the peak streamflow and the peak level of Missouri’s rivers are increasing over time, and that the severity of especially severe events is, too. I have heard this trend attributed to both levee building and climate change, but this data does not speak to causation.

Sources:

Sullivan, Brian K., Shruti Date Singh, and Mario Parker Bloomberg. 2019. “Hundreds of Barges Stalled as Floods Hider Midwest Supplies.” St. Louis Post Dispatch, 6/10/2019. Viewed online 6/10/2019 at https://www.stltoday.com/news/local/metro/hundreds-of-barges-stalled-as-floods-hinder-midwest-supplies/article_5a0355ea-3c03-584e-b3df-7a669205176d.html#tracking-source=home-top-story-2.

United States Geological Survey. National Water Information System: Mapper. I used the map to select the river gages for this article 6/10/2019 at https://maps.waterdata.usgs.gov/mapper/index.html.

United States Geological Survey. Peak Streamflow for the Nation. This is a data portal. I downloaded the data for the 9 river gages in this article on 6/10/2019 from https://nwis.waterdata.usgs.gov/usa/nwis/peak.

Births and Birth Rate Decline in 2018

Figure 1. Source: Hamilton, Martin, Osterman, and Rossen, 2019.

In 2018, 3,788,235 live births occurred in the United States, according to the Bureau of Vital Statistics. That is down from 3,855,500 in 2017, a decline of 2%. Figure 1 shows the trend in the data from 1990 to 2018. The number of births is the blue line, and it should be read against the left vertical axis. The fertility rate (the number of births per 1,000 women of child-bearing age) is shown as a green line, and it should be read against the right vertical axis.

The chart shows that both declined through 1997, then rose from 1997 to 2008, then began declining again. Overall, births have declined a little more than 10% since 1990. The fertility rate has declined by about 15% since then.

Every person in this world has an environmental footprint; we consume resources and create pollution. You can reduce the average environmental footprint, but you can’t eliminate it. The United States has the 7th highest per capita environmental footprint in the world, outranked only by Qatar, Luxembourg, the United Arab Emirates, Bahrain, Kuwait, and Trinidad and Tobago. Thus, the population of the United States is very important environmentally.

Figure 2. Data source: Missouri Information for Community Assessment.

According to the report, the number of births in Missouri in 2018 was 73,222. The report does not contain historical data for individual states, but data for 1990-2017 can be found on MICA, the Missouri Information for Community Assessment database. In Figure 2, the blue line represents the number of births, and it should be read against the left vertical axis. The MICA data does not include 2018, that has been added from the CDC report. The red line represents the fertility rate, and should be read against the right vertical axis.

The series are very similar, and they parallel the general shape of the data for the whole United States: the number of births and birth rate fell during the 1990s, then rose until around 2007, and have fallen since then.

Comparing the national fertility rate to that in Missouri, it appears that in 2017, the last year for which data was available in both jurisdictions, the national birth rate was just over 60 births per 1,000, while in Missouri it was 62. That doesn’t seem like a big difference, but multiplied over millions of people, it is substantial.

The report contains additional data regarding the race and ethnicity of the mothers giving birth, their age (teen births are a particular concern) and other characteristics.

Sources:

Global Footprint Network. Compare Countries. Data portal. Data for  2016 for “ecological footprint (gha per person).” Viewed online 6/7/2019 at http://data.footprintnetwork.org/#/compareCountries?type=EFCpc&cn=all&yr=2016.

Hamilton BE, Martin JA, Osterman MJK, Rossen LM. Births: Provisional data for 2018. Vital Statistics Rapid Release; no 7. Hyattsville, MD: National Center for Health Statistics. May 2019.

Missouri data generated 5/18/2019 on the Missouri Information for Community Assessment database (MICA): https://healthapps.dhss.mo.gov/MoPhims/MICAHome.

Fire and the Regeneration of Aspen Trees

Figure 1. Regeneration after the Red Eagle Fire in Glacier National Park. Photo: John May.

After returning from a trip to several national parks in 2016, I wrote a series of posts on wildfire, and the role wildfire has in keeping forests healthy. (See here.) In those posts, I reported that wildfire was essential for regenerating species of conifer that have serotenous cones. The cones of these species are coated with a waxy resin that prevents them from opening and releasing their seeds. Fire must melt the resin, and only then are the seeds released – millions of them. Thus, after a fire, the forest regenerates with thousands-upon-thousands of saplings, all the same age. Figure 1 shows the forest regenerating after the Red Eagle Fire near Glacier National Park. These are lodgepole pine, the dominant species in the forests of that area.

I also wrote that aspen trees require fire to regenerate. After a few decades, stands of aspen are invaded by conifers. Aspens are not shade tolerant, and they are not long-lived. Because the conifers create too much shade, the aspens cannot regenerate, and the stand dies out. Fire clears away the shade, and the aspen rhizomes, which remain beneath the ground, send up new shoots, and the aspen stand can be regenerated.

Figure 1. Effects of the Warm Fire (2006) in Kaibab National Forest. Photo by John May.

I just returned from the North Rim of the Grand Canyon. In 2006, the Warm Fire (what a name for a wildfire!) burned across Arizona Hwy. 67, the route to the North Rim. Figures 2, 3, and 4 show the scene. The Red Eagle Fire and the Warm Fire both occurred in 2006, but what has happened since is very different. The scene of the Red Eagle Fire is covered in thousands of small lodgepole pines, all the same age. The scene of the Warm Fire has nary a conifer to be seen. These are all aspens. They haven’t leafed-out yet, so they are a little difficult to see. Aspens turn brilliant colors in the fall – imagine what this area will look like when these trees are mature.

.

.

.

Figure 2. Effects of the Warm Fire (2006) in Kaibab National Forest. Photo by John May.

To my eye, the area burned by the Warm Fire looks blasted in a way that the area burned by the Red Eagle Fire does not. The reasons might include higher altitude, a more arid climate, and a hotter fire that sterilized the ground. But in addition, this is usually a mixed conifer forest. These species are less tolerant of full sunlight than are the aspens. Thus, the aspens recolonize the burned areas more quickly.

.

.

.

.

.

.

Figure 3. Effects of the Warm Fire (2006) in Kaibab National Forest. Photo by John May.

Eventually, an interesting thing will occur: the aspens will provide the light shade that the conifers need, and they will be able to start growing. In time, they will begin to shade out the aspens, which will die out, and there will be no more aspens until once again the area burns in a fire. Nature has her ways.

.

.

.

.

.

.

Figure 5. Warm Fire Progression Map. Source: United States Forest Service, 2006.

The Warm Fire was started in the Kaibab National Forest by lightning on 6/6/2006. At first, it was judged to be a small fire of low intensity that could be allowed to burn and would help renew the forest. In its first 10 days, it burned 1,049 acres.

After 2-1/2 weeks, however, suddenly it blew up into a very hot, rapidly-spreading fire. Between 6/23 and 7/4 it burned about 43,000 acres. Figure 5 shows the fire map through 6/27, but the fire wasn’t contained until 7/4.

Sources:

United States Forest Service. Warm Fire Recovery Project. Viewed online 5/27/2019 at https://www.fs.usda.gov/detail/kaibab/home/?cid=fsm91_050264.

United States Forest Service. Warm Fire Progression Map. Downloaded 5/27/2019 from https://www.fs.usda.gov/Internet/FSE_DOCUMENTS/fsm91_050152.pdf.

On a Break

Though I didn’t originally want to, I’m going to have to put MoGreenStats on a break until mid-June. I’m traveling, and while I hoped to be able to write as I travelled, I’m moving each day. I don’t have the chance to research an environmental report and write a post on it.

I will return home in June, and I should be able to have a new post prepared shortly after. Until then, be well.

 

No Decline in Missouri Crop Yields (Yet)

There have been some recent articles about how climate change is harming agriculture. One by Kim Severson in the New York Times (here) says “Drop a pin anywhere on a map of the United States and you’ll find disruption in the fields.” It goes on to discuss the impacts on “11 everyday foods”: tart cherries (Michigan), organic raspberries (New York), watermelons (Florida), chickpeas (Montana), wild blueberries (Maine), organic heirloom popcorn (Iowa), peaches (Georgia and South Carolina), organic apples (Washington), golden kiwi fruit (Texas), artichokes (California), and rice (Arkansas).

Well, that is a sampling of foods from around the country. I’m not so sure how “everyday” many of them are, but rice is certainly one of the basic grains.

A somewhat more convincing article by Chris McGreal in The Guardian interviewed farmers in valley of the Missouri River near Langdon, in northwestern Missouri. These are corn and soybean farmers. Their problem has been moisture: they have had too much rain. In many years, the ground has been so muddy that crops were ruined or not planted at all. In other years, the rain has caused the water table to rise so much that the ground looks dry on top, but is mucky mud just a few inches down. This is something, of course, that would affect river valleys the most, and the big river valleys in Missouri are some of the richest farmland the state has.

Figure 1. Data source: National Agriculture Statistics Service, USDA.

Most climate change studies project that climate change will impact agriculture negatively. Given this blog’s focus on the large statistical perspective, I thought it might be interesting to see how crop yields are doing in Missouri. The United States Department of Agriculture publishes the data. This data is a statistical average of yields across Missouri. Results in any one location may be different.

Figure 1 shows the per-acre yield for corn. The data shows that corn yields vary significantly from year-to-year, and that some years are really terrible, with yields being roughly half of what they are in good years. That said, there is a clear trend toward increased yields from 1957 right through 2014. Yields since then have been lower, and it is possible that we are looking at the start of a downward trend, but 4 years is not sufficient to tell.

Figure 2. Data source: National Agricultural Statistics Service, USDA.

Figure 2 shows the per-acre yield for soybeans. The yearly variation here may be somewhat less, but the overall pattern is much the same. With soybeans, however, yields increased right through 2017.

This data doesn’t tell us why crop yields are rising. Perhaps they are due to improved farming practices and better seed stock. It is possible that warmer temperatures, an increase in carbon dioxide, and more rain have benefitted crop yields overall, even if they have hurt some farmers in some locations. We just don’t know, at least not from this data.

What we do know is that, overall, the predicted negative effects of climate change do not yet seem to be reducing yields in these two important crops.

Sources:

McGreal, Chris. 2018. “As Climate Change Bites in America’s Midwest, Farmers Are Desperate to Ring the Alarm.” The Guardian,” 12/12/2018. Viewed online 5/1/2018 at https://www.theguardian.com/us-news/2018/dec/12/as-climate-change-bites-in-americas-midwest-farmers-are-desperate-to-ring-the-alarm.

Severson, Kim. 2019. “From Apples to Popcorn, Climate Change Is Altering the Foods America Grows.” The New York Times, 4/30/2019. Viewed online 5/1/2019 at https://www.nytimes.com/2019/04/30/dining/farming-climate-change.html?rref=collection%2Fsectioncollection%2Fclimate&action=click&contentCollection=climate&region=rank&module=package&version=highlights&contentPlacement=2&pgtype=sectionfront.

National Agriculture Statistics Service, United States Department of Agriculture. Quick Stats. This is a data portal that can be used to build a customized report. I focused on yield, in bushels per acre, for corn and soybeans from 1957-2018. Data downloaded 5/1/2019 from https://quickstats.nass.usda.gov.