Thursday, April 10, 2014

The Weather Forecasting Improvement Acts Passes the House: What are the Implications?

A most extraordinary thing has happened:  the U.S. House of Representatives has passed a bill that recognizes that the U.S. has fallen behind in weather prediction and proposes to do something about it.  The implications of this legislation are huge for the nation, NOAA, and the weather community.

Let me tell you the story.

On April 1st, the Weather Forecasting Improvement Act (H.R. 2413) was approved by voice vote in the U.S. House of Representatives and has now gone to the U.S. Senate. The principal sponsor is Congressman Bridenstine of Oklahoma, with 19 other sponsors, including some Democrats (such as Congresswoman Bonamici of Oregon).

There are some very good things in this bill, but some key problems were missed and the bill is far too detailed and prescriptive in many places.  The folks who wrote the bill were trying to do the right thing, but they clearly were not meteorologists and did not know where some of the skeletons were located in government closets.  So in this blog, I will go beyond critiquing what the bill does and tell you what the bill should have done, but didn't.

Let's face it, the existence of this bill is a real embarrassment for NOAA management. They have allowed weather prediction in the U.S. to degrade to such a obvious degree that Congress has felt the need not only to intervene, but to micromanage what NOAA/NWS should be doing to repair the situation.

The goal of the bill is stated up front:

To prioritize and redirect NOAA resources to a focused program of investment on affordable and attainable advances in observational,computing, and modeling capabilities to deliver substantial improvement in weather forecasting and prediction of high impact weather events, such as those associated with hurricanes, tornadoes, droughts, floods, storm surges, and wildfires and for other purposes.

And to make it perfectly clear about priorities:

...the Under Secretary shall prioritize weather-related activities, including the provision of improved weather data, forecasts, and warnings for the protection of life and property and the enhancement of the national economy, in all relevant line offices

The intention of this language is clear:  weather prediction should be the highest priority of NOAA

There are some folks that are uncomfortable with this language, arguing that weather and climate are interrelated and that climate research could be weakened.  Personally, I am sympathetic with this emphasis on weather prediction.  For too long, NOAA has  provided insufficient support for weather prediction, while dispensing far greater research and computational resources for climate studies.

Improved weather prediction can save lives and enhance economic activities now and in the future if the climate changes. There is no better protection from extreme weather that might accompany climate change than better weather prediction.

NOAA Bureaucracy:  Before I describe the rest of the bill keep in mind that NOAA (the National Oceanic and Atmospheric Administration) is an agency in the U.S. Department of Commerce.  The National Weather Service (NWS) is found within NOAA.   Atmospheric research is mainly NOT located in the National Weather Service, but rather in NOAA's Office of Oceanic and Atmospheric Research (OAR) and particularly in its Earth Sciences Research Lab (ESRL).   In this strange organization, the head of the National Weather Service or the head of operational weather modeling in the NWS  does not control the research supporting their operational mission.  As I will point out below, this has produced a dysfunctional and wasteful system, which Congress needs to change.

Back to the bill...

In the next section (3), the bill calls on the Assistant Administrator for NOAA OAR to conduct a research and development program to improve weather forecasting.  And the language is amazingly prescriptive (boundary layer processes, phased array technologies, cloud-resolving models, and more).   Quite honestly, too prescriptive.   And NWS  personnel are not involved in making the decisions of what research is to be done to SUPPORT THEIR MISSION.   Bad approach.

And then the bill does something wonderful: it calls for OAR to use not less than 30% of the allocated funds for extramural (outside of NOAA) research.   This is important.  In the past, NOAA has kept nearly all available research and development funds to use in-house, starving research in universities and the private sector.  A good example of this is the U.S. Weather Research Program (USWRP), which never succeeded in including the university community because NOAA would not supply sufficient funding.  Good for our Congressmen, this is one skeleton in the closest they knew about.  NOAA needs fresh ideas from the outside community and extramural research support is a key tool to secure it.

In the next two sections (4 and 5), the bill establishes a Tornado Warning Improvement and Extension Program and a Hurricane Warning Improvement Program.   You won't find much detail in these sections, except for a very aggressive goal of predicting tornadoes beyond one hour in advance.  Folks, we don't know if this is even possible.  And the Congressmen did not seem to know that we ALREADY have a hurricane improvement program on which they have spent tens of millions of dollars:  HFIP (the Hurricane Forecast Improvement Program).

Section 6 calls for Weather Research and Development Planning that requires that the Assistant Administrator for OAR will issue a research and development plan to restore and maintain U.S. leadership in numerical weather prediction and forecasting.   

Can't argue with that, except it is exceedingly strange that this is not being run by National Weather Service management.   As far as I can tell, there is no such plan in existence at this point.  Try to find a strategic plan for NWS numerical weather prediction over the next five years.  You can't, which is pretty amazing.  The Canadians have a detailed plan, so do the Brits.  We don't. 

In the next two sections (7 and 8), the Congressional gloves come off for an area in which NOAA has been dragging its feet for decades:  Observing System Planning

The U.S. spends billions of dollars on weather observations, and particularly satellite observations.   There are also weather radars, aircraft observations, surface observations, radar profilers, buoys, and much more.  Huge, expensive infrastructure.  You would think NOAA would design the observing system in a rational way to optimize quality of observations and reduce costs.  Think again.  They haven't.  And Congress is very frustrated about this.

There are technologies that are available to to such planning, one being Observing System Simulation Experiments (OSSEs) in which one simulates an artificial, but realistic, atmosphere and samples from it as if you were taking observations.  Then you use these observations in a simulated data assimilation and forecasting system.   At a hearing, two years ago, a Congressman asked the NOAA Administrator about this approach (not the current administrator) and she never even heard of OSSEs. Ouch.

Anyway, these sections demands the NOAA begin using OSSEs and other approaches to design a high quality and cost effective weather observing system.  It even cuts off funding for high-ticket observing systems unless they do it.   Good for Congress in pushing this issue and a real slap on the wrists of NOAA administrators. 

Section 9 of the act warms my heart.  It calls for Computing Resources Prioritization Report, which would would describe how NOAA would establish state-of-the-art computer facilities for weather prediction.  Congress is right. NOAA has been content to allow U.S. weather prediction facilities to deteriorate to a state that we have fallen out of numerical weather prediction leadership.  But Congress needs to do more than ask for reports and plans.  It needs to provide funds for supercomputers for weather prediction.  For the price of one jet fighter, the U.S. could improve weather prediction enough to save billions of dollars and protect American lives and property.   Real defense for Americans.  And remember Congress gave NOAA enough funds for a major computer and it hasn't been ordered yet because of NOAA's dependence on one vendor, IBM.

The next section (10), dealing with commercial weather data is a bit more controversial.  It gives NOAA a push toward taking more advantage of weather data provided by commercial providers.  NOAA already purchases some weather data from private companies (e.g., lightning data and some surface observations), but it needs to be more aggressive.  For example, NOAA is not purchasing extremely useful data from short-haul commercial flights (TAMDAR), data that is proven to improve short-term forecasts. It should.

But NOAA needs to be careful.  There are companies that are lobbying heavily to get NOAA to support their efforts when lower-cost options exist.  For example, some companies, like PlanetIQ, want to get into the GPS-occultation satellite business without investing in it themselves--looking for NOAA to front the investment by guaranteeing purchase of their data.  They are lobbying heavily to stop U.S. investment in the proven COSMIC system, for which other countries will pay at least half the cost.   Congress needs to learn the facts about such companies before they push NOAA to use their services.

Section 11 establishes a weather research and innovation advisory committee that will represent the national weather community, providing advice to NOAA on weather prediction technology and application.  A very good idea.  The NWS has resisted creating such a committee even though it has been recommended by many U.S. National Academy reports.  NOAA/NWS management clearly needs to be pushed on this.

Section 12, 13, 14:  Interagency Weather Research Coordination, Visiting Fellows at OAR and the NWS.   Certainly good ideas. 

This bill calls for expenditures of roughly 100 million dollars a year to fix problems, but does not provide new resources.  The question is where NOAA will secure the money:  reducing climate work, improving efficiencies, or some other approach?

My Take on This Legislation

As my analysis above indicates, there are many good aspects to this legislation, not the least of which is that it lets NOAA management know that inferior weather prediction for the U.S. in unacceptable and they need to make weather forecasting a national priority.

But unfortunately, this bill does not take on the true core problem in NOAA, the fact that the organization of  NOAA weather research and operations is basically flawed.  And until that is fixed, the U.S. will always lag behind in weather prediction

To move forward in weather prediction, research and operations must be integrated and coordinated.  This is what the highly successful European Center for Medium Range Forecasting does.   NOAA lacks this structure.   Operational numerical weather prediction is in the Environmental Modeling Center (EMC), a part of the National Weather Service.   Weather prediction research is mainly in NOAA's  ESRL lab, which is in NOAA OAR.   The head of EMC or his/her superiors in the NWS DO NOT CONTROL THE RESEARCH DONE TO SUPPORT THEIR MISSION.

The results of this structure has been a disaster for the country.  NOAA ESRL develops models that are never used by the National Weather Service, wasting substantial resources.  NOAA ESRL does some research that never has operational impact.  Don't get me wrong:  there are some extraordinarily talented scientists in NOAA ESRL and some of their work does get into operations (like the wonderful Rapid Refresh short-term forecasting system), but a huge amount of work goes no where. For example, NOAA ESRL is developing global models (NIM and FIM) that folks in the NWS assure me will never be applied operationally.

Boulder, Colorado  
Center of U.S. Weather Research

I got a taste of this a year and a half ago after I did a blog on improving the computer resources for the NWS.  The head of NOAA ESRL called me into his office and told me I should not work to get NWS better computers; rather,  I should work to secure them for NOAA ESRL!  That we would set up an essential real-time forecasting capability separate from EMC.  And he made deprecatory comments about his NWS colleague's technical abilities.  Folks, this is a very broken system.

So what should Congress do?   They should combine U.S. government operational numerical weather prediction and research into one entity.  Specifically, they should combine:
  • The NWS Environmental Modeling Center (EMC)
  • The NOAA ESRL Global Systems Division
  • The NWS Meteorological Development Lab (MDL)
into one entity, headed by one person.   MDL does model post-processing (statistically improving model output).   I suspect the best place for this combined unit would be in Boulder, Colorado:  NOAA ESRL is there already and Boulder is the home of the National Center for Atmospheric Research (NCAR) .  In fact, some portions of NCAR, like the Developmental Testbed Center, could be entrained into the new national forecasting entity. Boulder is the center of U.S. weather research and also a very pleasant location that is attractive to visitors.

The creation of such a new entity, the provision of large supercomputer resources,  ensuring sufficient funds for extramural, weather-prediction related research , and developing an approach for coordinating the huge research/development weather capabilities of the U.S. would rapidly improve U.S. weather prediction in NOAA.  But only Congress can make this happen...NOAA management never will.  And these issues extend beyond NOAA:  there is a huge amount of weather research and operations in other agencies, including  the U.S. Navy, U.S. Air Force, U.S. Department of Interior, NASA and elsewhere.   There is considerable redundancy, duplication of effort, and lack of coordination between these weather forecasting groups.  In the end, the problem is not resources, its lack of coordination and cooperation.  And we haven't even talked about the capabilities of the U.S. private sector.  What a mess.   And only Congress or the leadership of the U.S. Executive Branch can fix it.

The legislation now goes to the Senate and I hope they will make some of the above suggested changes in the bill.


Beth Niquette said...

Like, this will do any good, right? lol Micromanaging by our lawmakers seems the way of things these days.

Unknown said...

All national weather forecasting under one roof. Weather research under NWS control. This seems so elementary.

Great report Dr. Mass. I certainly have a better understanding of this government problem.

Fred said...

the existence of this bill is a real embarrassment for NOAA management

I'd rephrase this to say "NOAA management is a real embarrassment [to] NOAA management"

As you mentioned, as long as OAR and the NWS are separate line offices, the same old bureaucratic turf wars will continue and little will change. NOAA desparately needs reorganization, but that would require courageous leadership. And you don't get to the top of the NOAA by being bold.

DC Collyer said...

If you are interested the East Coast of Australia is about to be hit by a superstorm, follow the link below for more info.I live in the cyclone Ita affected area, so i decided to start a blog video journaling the effects of the cyclone. I am hoping & crossing my fingers my home remains intact. My Blog Is here if you want the latest from a local perspective!

Michael Snyder said...

For me a its a victory that our congress even pays any attention to the details of NOAA. This bodes well for the future I believe.

WxAnalyst said...

This is just smoke and mirrors. There are significant parts of weather activity within DOC(NOAA), USDA, DOI, EPA, the military and more. If the government had a clue, they would all be consolidated to improve weather and climate prediction and save money at the same time. Look at the number of redundant climate offices opening; look at Congress putting pieces of weather activity in every state of the union and you see that Congress is not the solution. It is the problem. Consolidate all government weather/water/ climate and environmental activities in one place. It's the only real answer.

Randy Peppler said...

Dr. Mass, you say that weather research in NOAA is primarily done at ESRL (and CIRES). This is a bit of a stretch - you need to include NSSL/CIMMS and also CIRA here. Other labs/CIs also do some weather research.

Brad said...

What about the "O" in NOAA? As in, the Ocean? "The Under Secretary shall prioritize weather-related activities". Doesn't leave much room for the most important fluid on the surface of the Earth (and air, my fellow atmos sci academic, ain't it!)

Cliff Mass said...

I am talking about research on Numerical Weather Prediction. Most numerical weather prediction research in NOAA is done in ESRL...cliff