California drought

In the September 2014 issue of the Bulletin of the American Meteorological Society (BAMS), there was a compilation of papers published entitled "EXPLAINING EXTREME EVENTS OF 2013 FROM A CLIMATE PERSPECTIVE." The California drought was among the extreme events looked at in that series of studies.

One study (Swain, et al.) suggested that upper air flow circulations that persisted over that region were influenced by AGW, but did not make an explicit attribution claim regarding extremely low precipitation. Two other studies found no clear link to AGW (Funk, et al., and Wang and Schubert).

Actually, the only clear links to AGW in the BAMS studies were heat waves in various areas around the world (5 extreme heat waves). Extreme precipitation events of 2013 were found to have been much less influenced by human-induced climate change than extreme high temperature events, and "stormy" weather events (e.g., blizzards, tropical cyclones) were not found to have any linkage to AGW.

 

Funk et all !
Texas drought was supposed to be caused by global warming .Now the floods are blamed on global warming.

But historically Texas flood and drought was the same severity as today.
Everything bad today is caused by global warming. The floods and droughts of the past were caused by witches.

 

#1 PDSI uses both known temperature and precipitation, and some empirical equations to turn those into soil water balance. It is what engineers might refer to as a kludge, but has become entrenched in the field. Also it was originally intended as a agriculture tool and has now been extended to all vegetation types.

Whether PDSI or SPI are a better fit to a particular set of tree-ring data - well it depends I guess. But some reliance on proxies is necessary if we are to look beyond the thermometers and rain buckets.

It might not be common knowledge, but the Big Daddy of tree-ring research was after one thing above all. That was tree-growth evidence for the 11-year solar cycle! He found it, too, but also found much more variations that eventually led to the entire science of paleoclimatology.

Wood density, as well as ring width can be extracted from tree corings, and I don't know know whether PDSI or SPI generally fit better there.

I fully expect that the outermost 4 or 5 growth rings in Calif. trees would be a narrow set compared with anything earlier in the regional record. That spans 2000 years in some forest types and much shorter in others. So, a fella might say "it has been xxx long since trees grew so slowly'. Thus avoiding any climatological interpretation. Just tree vs. tree. Send the manuscript off to a journal and I'm quite sure the reviewers will ask 'why no PDSI or SPI?' This is how 'entrenched' happens.

Whether or not Calif. drought is strongly linked to current climatic conditions. Current water shortage is a fact and the affected areas ought to be preparing for the next one.

If El Nino bulks up, there will probably be good snowpack in west coast mountains (their southern half) next winter. Before that, more drought, I would think.

 

The idea of a 100-year flood (for example) is derived from statistical analyses that have been traditionally been presented as exceedance graphs. One or both axes may be logarithmic, or the return intervals presented in even more arcane units. The point is, once you have a history of that local data, any event can be described according to its mean return interval.

If a location had two big floods, the time between them can be compared to the underlying distribution, and a measure of the (im) probability of the situation follows.

It is the standard procedure to describe streamflows, but I don't see the technique applied to temperature (extremes) or rainfall, or length of periods without rain (that lead into drought).

Incidentally, nothing precludes two, 100-year floods in two successive years. But its probability (0.00tiny) can be defined

 

"Wouldn't a 100 year flood have a 1% of happening in any given year

That is a precise definition! In no way does it EXCLUDE100-y floods happening back to back. Rather it allows the calculation of probability that such a thing could happen.

AG also suggests separating ENSO + and - years in probability distributions and graphical analyses. I think this is a SUPERB idea that should be brought to the attention of hydrologists at UTA or elsewhere. The high-end journals are hungry for new insights from old data. Houston and other parts of TX are still sloppy wet. CA is unusually dry. Do it, my friend. I won't steal your idea (pinned down by wood decomposition). Other PC readers won't either, as they can't be bothered to read 1 (or 30) published papers.

It's all you.

 

"I would think soil moisture ... would be important to analysing how severe a drought or a flood would be."

The PDSI is supposed to display soil moisture, but it is inadequate because of relying on indirect measures and empirical blah blah blah. Anyone can measure soil moisture anywhere (everywhere) with neutron probes or time-domain reflectometers. Those are very expensive. Alternative is to take (thousands of) soil samples, weigh, dry, weigh again. That would work well if you have large numbers of students, and you use tests to exclude their crappy data. They are children, and generate crappy data at some level of incidence.

Drought is drought when plants cannot take enough water from soil to support photosynthesis. This depends on soil-water content (as above) but also on the particle-size distributions of soils. See, when you open the box, you find lots of other boxes inside.

The other extreme is flood. Flood means that current water-input rates exceed the saturated hydraulic conductivity of soils, so there is surface runoff (Hortonian flow ). To understand this at the micro level requires a lot of local sampling.

While I should not dissuade AG from addressing either extreme, the work required is daunting. I see more hope from moving to larger scales where ENSO +/- seems important.

Data in US for local rainfall and stream flow are on the 100-yr scale at best. This is not many ENSO cycles; too bad about that, sorry. But in space there are thousands of watersheds (catchment basins, places where input water gets concentrated into output water). In time, we are f***ed. In space, things look pretty good. Not all strong retrospective hydrology research has yet been published

 

California drought: Past dry periods have lasted more than 200 years, scientists say - San Jose Mercury News

" "Through studies of tree rings, sediment and other natural evidence, researchers have documented multiple droughts in California that lasted 10 or 20 years in a row during the past 1,000 years -- compared to the mere three-year duration of the current dry spell. The two most severe megadroughts make the Dust Bowl of the 1930s look tame: a 240-year-long drought that started in 850 and, 50 years after the conclusion of that one, another that stretched at least 180 years. "

 

Not that I think that CO2 has anything to do with it.But that graph above coincides pretty well with the medieval warm period.
I hope its not the recovery from the Little Ice Age causing drought in the west.Because we have recovered.

 

That Cook study shows pretty much 50% drought over the past 1200 years.
600 years of drought.600 years wet.
With those kind of odds ,Id say we're pretty screwed .
Except that Earth is predicted to be entering a 30 year period of cooling . Ca. may be saved by the Sun and ocean currents.
GAIA wont burn all the California "down to Earth" types.

 

Defined based on what, the totally unrelated 100 year flood number?

Mathematically, that comment is based upon an assumption of randomness that probably does not exist. Rather than delve into the mathematics, let me give an example. Let's say 100 year floods alway occur in runs of 10 to 300 years but over 1000000 years, only 10000 floods have occurred. So the flood probability is 1 per 100 years but the probability of a 100 year flood being proceeded or followed with another flood is 100%. You needed to clearly state the "definition" of the successive probability is completely different that the 1 year probability.

 

Cook and the LDEO are definitely doing the most with drought reconstructions (above graph based on 2007 publication). Their newer work also deserves our attention

http://advances.sciencemag.org/content/1/1/e1400082

21 MB NOAA report on CA drought can be downloaded here


Climate Program Office > Climate Programs > Modeling Analysis Predictions and Projections > MAPP Task Forces > Drought Task Force > California Drought

 

...

 

FL, the assumptions are independence of events and stationarity. That's just the way the 'flood folks' work. The short (century-scale) stream flow records are a substantial limitation on the wet side.

On the dry side, paleo proxies such as those used by Cook et al to create the North America Drought Atlas have enough time span so that (assuming proxies are measuring the desired thing) , then independence and stationarity can be tested.

The graph in #11 argues against stationarity. Rather that there exist long-term climate regimes.

Many moons ago, the monthly newsletter of a famous technical university in Pasadena CA had an article about nuclear power safety. The notable message was along the lines of "xxx happens every 20 years, it just happened , so we are safe for the next 20."

a rather embarrassing example of getting such things wrong.

 

But what actually makes those assumptions valid?

This is not criticism, but just something fun to point out. The truth of the issue is all floods are completely deterministic...but not necessarily determinable by man. So we resort to probabilities to characterize this lack of real understanding. Then we soon fall into the trap of thinking nature is operating according to probabilities. It is not. It just seem like it until a better understanding is reached.

So in the discussion of the draught and many other climate subjects, I see this happen. All the draughts had mechanisms causing them. Unfortunately, a lot of the AGW approach is to "sort of, kind of, maybe possibly" have a probability link of affecting the draught. This then leads to trying to find out the mechanisms by seeing what happens to probabilities due to running models instead of trying to find the mechanisms actually in play. The end result can be a highly praised model where the statistics of the model match the statistics of the draughts....but the model could have all the internal mechanisms wrong....leading to a dead end.

 

I have not before here discussed my limited personal reliance on statistics. Not that I can recall anyway.

We lack mechanistic understanding of global hydrology, and when one scales down to a local flood plain, it only gets worse. But one must do something, so recurrence analyses fill the void.

If there are strong, ocean-mediated climate regimes (see Tsonis), and / or if high infrared absorption can move us to a different regime (see 'the 97%'), then stationarity-based statistics offer less utility.

Of course, there are climate models (my first link @15) but then it's a balance between faith in Cook's research (expressed by mojo @11) and less faith in models' ability to get hydrology right (expressed by me and AustinG and lots others).

But, there they are. Projections concerning hydrology in agric areas and dense population areas might lead to more rational water plans. In contrast to 'mainstream' models, dissenters have offered, well, nothing that I can discern.

Which leads us back to 30-year cooling (again), so I'll initiate a thread on that, that mojo chose not to do.

 

But in reality scientists are paid to say that theres doom and gloom if society continues to be productive.
Take this guy Cook for example. Great at gathering data .
But a grant whore when he has to analyze the data.
$100 billion dollars in US funding, I dont blame them one bit.