Great piece Roger, and I am glad you included the quotes from me in 2011. I have two thoughts to add regarding floods. You mention several reasons the flood frequency distributions can change over time. One of those is urbanization, which you mention. The data, for small watersheds shows massive amounts of change over a few decades as the transition is made from rural to urban. Another big driver is artificial land drainage (e.g. tile drainage)which is designed to get the water off the landscape, and of course where it ends up is in the river, resulting in bigger floods. This practice is common in the upper Midwest and has a notable impact. The other thing I want to mention is that this situation of very large shifts in flooding at time scales of half a century or more is at its most pronounced in the area of the Red River of the North. Specifically the eastern parts of North and South Dakota plus parts of Minnesota and Iowa. This is partly a consequence of this region going back and forth between being semi-arid and humid and getting “stuck” in one of those states for decades. That is related to soil types, topography, and some kind of quasi-periodic climate oscillations. My point here is that this region is a poster child for the kind of confusion between trend and variability is so common in the world of hydrologic extremes. The problem exists everywhere but is at its most dramatic in this region. Thanks for the great discussion.
Again a very interesting and thought provoking post. I understand that your message is more about what constitutes "current climate" and how hard it is to actually define that.
I would like to point to something that you have addressed partly before, but it might be worth diving into. It is the Bell curve in the EPA figure you show. You state, correctly, that it is deeply misleading. You then address the "Current climate" part of it, since this is the focus of your post.
The figure shows a situation where the mean of climate shifts, but the distribution does not change shape - i.e. the relative extremes stay the same. It does not become wider or more narrow. This may be a moot point since the normal distribution probably is not the right one to use. I think however that there is something to look into here.
In my country, Norway, the story in the late 1990s/early 2000s used to be that we were looking towards a future where climate would become warmer, wetter and wilder. The "wilder" bit of this was linked to wind. Since then the statistics have shown that it is indeed getting warmer - mainly in winter - and we have more precipitation. But wind speeds have not changed. If anything, they are lower - as would be the case if indeed the Arctic is warming faster than the rest of the world. Less temperature gradient gives less potential for wind.
So they have gradually changed the narrative for "wilder" to mean more temperature hikes and more extreme precipitation events. The temperature part is a hard sell as we see higher maximum temperatures but not that many abrupt "wild" changes. With precipitation we actually do not have numbers on such events as they are very local and most are not caught by measurement infrastructure. Instead of recognizing this and saying that we just don't know, the do the climate science thing and say absence of evidence is not evidence of absence. To me it seems more about supporting the narrative than it is about actually understanding what is going on.
Anyway, the issue of extremes is a very interesting one and especially how it shifts if and when "current climate" changes.
I have a book I purchased a few years back wherein Bob Tisdale downloaded from NOAA a bunch of data including information on temperature, precipitation and drought from 1918/2018, meaning that I probably bought the book in 2019. For temperature, he chose, of all the data available, to present the highest (and lowest and average) monthly temperature each month during that 100-year period. He listed the data by state, but since political boundaries don't tell us much, he also listed it by larger areas of the country. Nonetheless, for most states of the lower 48, I noticed a spike in temperatures in the 1930s higher than anything since. In my home state of Colorado, Bob recorded a 100-year trend of +.156°F per decade, or a 100-year trend of +1.56°F. That is barely perceptible, and I daresay may not be out of line with the long-term trend since the end of the Little Ice Age. This is the problem I have with CO2 as the control knob - unless it is U.N. Secretary-General Antonio Guterres warning us that the planet is "boiling", it appears to me that Climate Change is a giant perception management affair.
A few years ago I learned that conventional pulmonary function tests (e.g., FEV1 [forced expiratory volume in 1-sec]) have, since the 1970s, been conducted in a way that doesn’t collect enough data and discards most of the data collected. We simulated intra-test variability and discovered that typical
comparisons between baseline and pollutant-exposed FEV1 overstated the difference resulting from exposure.
See Belzer, Richard B. and R. Jeffrey Lewis. 2019. “The Practical Significance of Measurement Error in Pulmonary Function Testing Conducted in Research Settings,” Risk Analysis, 39:10 (2316-2328).
I believe (but have not followed up) that conventional blood pressure data have a similar problem. Probably a lot of other biomedical data we take for granted as reliable but isn’t.
About 15 years ago I had an enterprise in Cedar Rapids, Iowa. We rented an office well above the level of the Cedar River. The locals assured us that the river never flooded. Further I checked the flood maps and our office was in the 500 year flood zone. Within a short time there was 8 feet of water in the building.
Roger you made a few references to changes in climate being caused by CO2. Obviously this a measure that is increasing and measurable. But it not the only human influence on the environment and therefore climate. Increasing flood activity might more likely be the cause of land use change causing more water to enter the waterways. It may in fact have nothing to do with clime change nor rising CO2.
The narrow focus detracts from the totality of human intervention or effect.
If attribution of long-term trends in climate-related variables to greenhouse-gas-driven climate change is impossible because of the non-stationarity of observed distributions, then, by the same reasoning , the absence of long-term trends is not persuasive evidence of there being no effect from greenhouse-gas-driven climate change. In other words, just as correlation is not causation. the absence of evidence is not to be confused with evidence of absence. So are we left with the deeply unsatisfying impression that maybe we shouldn’t be trying to evaluate climate variables ? I think not. Climate science is basically an observational science that is still developing it’s theoretical underpinnings.
We do know from solid experimental evidence (spectroscopy) backed with a solid theoretical underpinning (radiative transfer) that CO2 and other gases cause warming. And we also know that humans are causing at least some of that increase in the atmospheric burden of those gases. So we humans have a direct a responsibility for adding some amount of heat to the system. I maintain that we have an obligation to understand the consequences of our actions. No, we can’t yet predict very well how that heat will distribute nor manifest itself through the atmosphere/ocean/terrestrial/biosphere systems. But it’s incumbent upon us to keep sifting through the imperfect evidence for clues and insights. Else the theoretical understanding of the underlying atmospheric physics and chemistry and will never improve. And we humans will never understand the consequences of our actions nor be able to formulate prudent policies of "no regrets" to address them.
The change from midnight to midday also induces warming of the atmosphere. The atmosphere then sheds some of the warming, then rinse and repeat. Nobody knows if this daily heat accumulates in the atmosphere, or is shed.
If CO2 in the atmosphere induces some warming, the atmosphere sheds some of the warming, rinse and repeat.
Why single out CO2 for special treatment, leading to expensive and drastic policies about hydrocarbon fuels?
Nobody knows if CO2-generated heat accumulates in the atmosphere or is shed.
It's also encumbent upon humans to evaluate the total economic impact on humans and the environment of any mitigation measures. As the COVID response demonstrated, optimizing (e.g., lockdowns)for a single variable of a complex multi-variate nonlinear system had deleterious effects on other variables affecting mankind (e.g., learning loss, increased substance abuse, suicide rates, etc.)
More colloquially, the juice has to be worth the squeeze.
I have long felt that the importance of long records for the water level of the Nile River needs more expression in the climate change context. (I am now too old for some of this).
Roger, sorry to double up on the comments, but after I posted the comment below I had a brain storm which might be worth pursuing. There are a very few temperature and weather records of very long duration. Surely these should be studied for the possibility that they might help define natural variability on very long time scales? I understand that records of the height of the Nile River in flood were kept from ancient times up until the Aswan Dam interrupted the natural rhythm of the river. The Central England Temperature record goes back, if memory serves, to the middle 1600's, almost 400 years now. Records of the dates of first cherry blossoms in Japan similarly go back many hundreds of years, and I would be amazed if Chinese records could not yield something similar going back long before mercury thermometers in Stevenson screens. Just a thought.
Roger, this is excellent. The whole idea of using a bell curve to philosophize about temperature assumes that temperature follows a Gaussian distribution. Since weather is a chaotic, non-linear phenomenon, affected by a hodge-podge of natural cycles of differing cycle length, assuming Gaussian statistics for a variety of weather phenomena seems like a dubious and unproven assumption. It would seem that extreme value statistics would be more appropriate. Perhaps the mathematicians here at THB can give us some education on this topic.
Spot on. Gaussian distribution is great for coin tosses and dice rolls. Leptokurtic distributions are better, but in the end, highly complex systems are impossible to model beyond a very short time horizon.
As a sailor, I ignore weather forecasts beyond 72 hours, and put a heavy weighting on those within 24 hours.
What is apparent to the even casual observer is that the geologic history of the planet, billions of years for all most all of which man did not exist, has been written by extremely powerful NATURAL forces. Some of these we understand and try and measure, but our measurements are only available for an exceedingly brief period of time and of limited use. These natural forces are generally very slow-moving take place over millions of years when compared to a short human life span. It's just hard to believe that any serious or honest person could think that mankind's contribution to atmospheric CO2 would impact the planet's climate in a meaningful way.
Activists like to point to shrinking glaciers as a sign of anthropogenic global warming (climate change).
In the 1790s Captain George Vancouver was sailing up the Pacific Coast, mapping the coast line, when he stopped at the entrance to Alaska's Glacier Bay. It wasn't much of a bay - more like an indentation in the coast line. The glacier extended almost to the ocean.
Eighty some years later, in 1879, John Muir (founder of the Sierra Club) while exploring in the area found that the glacier had retreated about 45 miles since Vancouver's visit. Maximum glaciation in the area was about 1750, when the glacier was a thumb, sticking out past the shore for 5 miles.
So, the glacier retreated 50 miles from 1750 to 1879, when humans could have had nothing to do with the shrinkage. NOAA likes to use 30 years for long term climate averages, when in this case, the 1,000 year Eddy solar cycle would be more appropriate.
Excellent post followed by excellent comments! When people attempt to show that some undesirable aspect of extreme weather is increasing I can't help but pose the question (to myself) as to how likely a such a small increase in temperature (~1.5°C?) could be the cause. I may be missing something basic here but greenhouse gases cause a temperature increase and everything else has to stem from that if we're to blame emissions. We know roughly what the temperature increase has been irrespective of the cause and it's not much.
Great piece Roger, and I am glad you included the quotes from me in 2011. I have two thoughts to add regarding floods. You mention several reasons the flood frequency distributions can change over time. One of those is urbanization, which you mention. The data, for small watersheds shows massive amounts of change over a few decades as the transition is made from rural to urban. Another big driver is artificial land drainage (e.g. tile drainage)which is designed to get the water off the landscape, and of course where it ends up is in the river, resulting in bigger floods. This practice is common in the upper Midwest and has a notable impact. The other thing I want to mention is that this situation of very large shifts in flooding at time scales of half a century or more is at its most pronounced in the area of the Red River of the North. Specifically the eastern parts of North and South Dakota plus parts of Minnesota and Iowa. This is partly a consequence of this region going back and forth between being semi-arid and humid and getting “stuck” in one of those states for decades. That is related to soil types, topography, and some kind of quasi-periodic climate oscillations. My point here is that this region is a poster child for the kind of confusion between trend and variability is so common in the world of hydrologic extremes. The problem exists everywhere but is at its most dramatic in this region. Thanks for the great discussion.
Superb piece, accompanied by insightful comments. What a pleasure.
Again a very interesting and thought provoking post. I understand that your message is more about what constitutes "current climate" and how hard it is to actually define that.
I would like to point to something that you have addressed partly before, but it might be worth diving into. It is the Bell curve in the EPA figure you show. You state, correctly, that it is deeply misleading. You then address the "Current climate" part of it, since this is the focus of your post.
The figure shows a situation where the mean of climate shifts, but the distribution does not change shape - i.e. the relative extremes stay the same. It does not become wider or more narrow. This may be a moot point since the normal distribution probably is not the right one to use. I think however that there is something to look into here.
In my country, Norway, the story in the late 1990s/early 2000s used to be that we were looking towards a future where climate would become warmer, wetter and wilder. The "wilder" bit of this was linked to wind. Since then the statistics have shown that it is indeed getting warmer - mainly in winter - and we have more precipitation. But wind speeds have not changed. If anything, they are lower - as would be the case if indeed the Arctic is warming faster than the rest of the world. Less temperature gradient gives less potential for wind.
So they have gradually changed the narrative for "wilder" to mean more temperature hikes and more extreme precipitation events. The temperature part is a hard sell as we see higher maximum temperatures but not that many abrupt "wild" changes. With precipitation we actually do not have numbers on such events as they are very local and most are not caught by measurement infrastructure. Instead of recognizing this and saying that we just don't know, the do the climate science thing and say absence of evidence is not evidence of absence. To me it seems more about supporting the narrative than it is about actually understanding what is going on.
Anyway, the issue of extremes is a very interesting one and especially how it shifts if and when "current climate" changes.
I have a book I purchased a few years back wherein Bob Tisdale downloaded from NOAA a bunch of data including information on temperature, precipitation and drought from 1918/2018, meaning that I probably bought the book in 2019. For temperature, he chose, of all the data available, to present the highest (and lowest and average) monthly temperature each month during that 100-year period. He listed the data by state, but since political boundaries don't tell us much, he also listed it by larger areas of the country. Nonetheless, for most states of the lower 48, I noticed a spike in temperatures in the 1930s higher than anything since. In my home state of Colorado, Bob recorded a 100-year trend of +.156°F per decade, or a 100-year trend of +1.56°F. That is barely perceptible, and I daresay may not be out of line with the long-term trend since the end of the Little Ice Age. This is the problem I have with CO2 as the control knob - unless it is U.N. Secretary-General Antonio Guterres warning us that the planet is "boiling", it appears to me that Climate Change is a giant perception management affair.
The comments were as great as the post today. A little over my head but terribly ensightful.
A few years ago I learned that conventional pulmonary function tests (e.g., FEV1 [forced expiratory volume in 1-sec]) have, since the 1970s, been conducted in a way that doesn’t collect enough data and discards most of the data collected. We simulated intra-test variability and discovered that typical
comparisons between baseline and pollutant-exposed FEV1 overstated the difference resulting from exposure.
See Belzer, Richard B. and R. Jeffrey Lewis. 2019. “The Practical Significance of Measurement Error in Pulmonary Function Testing Conducted in Research Settings,” Risk Analysis, 39:10 (2316-2328).
I believe (but have not followed up) that conventional blood pressure data have a similar problem. Probably a lot of other biomedical data we take for granted as reliable but isn’t.
About 15 years ago I had an enterprise in Cedar Rapids, Iowa. We rented an office well above the level of the Cedar River. The locals assured us that the river never flooded. Further I checked the flood maps and our office was in the 500 year flood zone. Within a short time there was 8 feet of water in the building.
Roger you made a few references to changes in climate being caused by CO2. Obviously this a measure that is increasing and measurable. But it not the only human influence on the environment and therefore climate. Increasing flood activity might more likely be the cause of land use change causing more water to enter the waterways. It may in fact have nothing to do with clime change nor rising CO2.
The narrow focus detracts from the totality of human intervention or effect.
If attribution of long-term trends in climate-related variables to greenhouse-gas-driven climate change is impossible because of the non-stationarity of observed distributions, then, by the same reasoning , the absence of long-term trends is not persuasive evidence of there being no effect from greenhouse-gas-driven climate change. In other words, just as correlation is not causation. the absence of evidence is not to be confused with evidence of absence. So are we left with the deeply unsatisfying impression that maybe we shouldn’t be trying to evaluate climate variables ? I think not. Climate science is basically an observational science that is still developing it’s theoretical underpinnings.
We do know from solid experimental evidence (spectroscopy) backed with a solid theoretical underpinning (radiative transfer) that CO2 and other gases cause warming. And we also know that humans are causing at least some of that increase in the atmospheric burden of those gases. So we humans have a direct a responsibility for adding some amount of heat to the system. I maintain that we have an obligation to understand the consequences of our actions. No, we can’t yet predict very well how that heat will distribute nor manifest itself through the atmosphere/ocean/terrestrial/biosphere systems. But it’s incumbent upon us to keep sifting through the imperfect evidence for clues and insights. Else the theoretical understanding of the underlying atmospheric physics and chemistry and will never improve. And we humans will never understand the consequences of our actions nor be able to formulate prudent policies of "no regrets" to address them.
Gene M,
The change from midnight to midday also induces warming of the atmosphere. The atmosphere then sheds some of the warming, then rinse and repeat. Nobody knows if this daily heat accumulates in the atmosphere, or is shed.
If CO2 in the atmosphere induces some warming, the atmosphere sheds some of the warming, rinse and repeat.
Why single out CO2 for special treatment, leading to expensive and drastic policies about hydrocarbon fuels?
Nobody knows if CO2-generated heat accumulates in the atmosphere or is shed.
Geoff S
It's also encumbent upon humans to evaluate the total economic impact on humans and the environment of any mitigation measures. As the COVID response demonstrated, optimizing (e.g., lockdowns)for a single variable of a complex multi-variate nonlinear system had deleterious effects on other variables affecting mankind (e.g., learning loss, increased substance abuse, suicide rates, etc.)
More colloquially, the juice has to be worth the squeeze.
I have long felt that the importance of long records for the water level of the Nile River needs more expression in the climate change context. (I am now too old for some of this).
Geoff S
Funny, I was just watching a YouTube on the ancient Pharaohs and it talked about a 50 year drought that devastated the country and region.
I watched some on Angkor Wat. A hydro engineering marvel that was abandoned after climate change reduced its rainfall.
Roger, sorry to double up on the comments, but after I posted the comment below I had a brain storm which might be worth pursuing. There are a very few temperature and weather records of very long duration. Surely these should be studied for the possibility that they might help define natural variability on very long time scales? I understand that records of the height of the Nile River in flood were kept from ancient times up until the Aswan Dam interrupted the natural rhythm of the river. The Central England Temperature record goes back, if memory serves, to the middle 1600's, almost 400 years now. Records of the dates of first cherry blossoms in Japan similarly go back many hundreds of years, and I would be amazed if Chinese records could not yield something similar going back long before mercury thermometers in Stevenson screens. Just a thought.
Roger, this is excellent. The whole idea of using a bell curve to philosophize about temperature assumes that temperature follows a Gaussian distribution. Since weather is a chaotic, non-linear phenomenon, affected by a hodge-podge of natural cycles of differing cycle length, assuming Gaussian statistics for a variety of weather phenomena seems like a dubious and unproven assumption. It would seem that extreme value statistics would be more appropriate. Perhaps the mathematicians here at THB can give us some education on this topic.
Spot on. Gaussian distribution is great for coin tosses and dice rolls. Leptokurtic distributions are better, but in the end, highly complex systems are impossible to model beyond a very short time horizon.
As a sailor, I ignore weather forecasts beyond 72 hours, and put a heavy weighting on those within 24 hours.
What is apparent to the even casual observer is that the geologic history of the planet, billions of years for all most all of which man did not exist, has been written by extremely powerful NATURAL forces. Some of these we understand and try and measure, but our measurements are only available for an exceedingly brief period of time and of limited use. These natural forces are generally very slow-moving take place over millions of years when compared to a short human life span. It's just hard to believe that any serious or honest person could think that mankind's contribution to atmospheric CO2 would impact the planet's climate in a meaningful way.
Another example:
Activists like to point to shrinking glaciers as a sign of anthropogenic global warming (climate change).
In the 1790s Captain George Vancouver was sailing up the Pacific Coast, mapping the coast line, when he stopped at the entrance to Alaska's Glacier Bay. It wasn't much of a bay - more like an indentation in the coast line. The glacier extended almost to the ocean.
Eighty some years later, in 1879, John Muir (founder of the Sierra Club) while exploring in the area found that the glacier had retreated about 45 miles since Vancouver's visit. Maximum glaciation in the area was about 1750, when the glacier was a thumb, sticking out past the shore for 5 miles.
So, the glacier retreated 50 miles from 1750 to 1879, when humans could have had nothing to do with the shrinkage. NOAA likes to use 30 years for long term climate averages, when in this case, the 1,000 year Eddy solar cycle would be more appropriate.
Does this mean the planets not going to catch on fire in the next five years?🤔
Excellent post followed by excellent comments! When people attempt to show that some undesirable aspect of extreme weather is increasing I can't help but pose the question (to myself) as to how likely a such a small increase in temperature (~1.5°C?) could be the cause. I may be missing something basic here but greenhouse gases cause a temperature increase and everything else has to stem from that if we're to blame emissions. We know roughly what the temperature increase has been irrespective of the cause and it's not much.