climate change, Science

Adapting to a new world

The Paris Climate Agreement provides us with a historic and far reaching consensus that will help us adapt to a changing world. It could literally save our lush, productive yet fragile planet.  The agreement rests on its “shall” and “should” tasks.  The difference is important.  Imagine your choking on a chicken bone and your wife says, “I should give you the heimlich maneuver”.  Shall would be more appropriate.  In the agreement the “shells” would get us to limiting the world temperature increase to 2 degrees C (3.5 degrees F).  This would stop the climate from going off the rails.


sunset over a marsh

sunset over egg island, nj


But more exciting for conservationists are the “shoulds”.  If implemented our world would stay to within 1.5 degrees C , a threshold that most scientists agree would protect us from serious changes.   To do that we will need all hands on deck from renewable energy to carbon sequestration offsets.  The world would indeed change, but in a way that would allow my generation to hand off a world in good shape to the next generation. At least it will be up to them.

But for wildlife biologists and naturalists the agreement points to an emerging perspective that is, in itself, is a bold undertaking.  The agreement rests on scientific evaluation to inform the progress towards the various tasks of limiting carbon emissions in enough time to make changes to the tasks.   One could say this idea, informing policy through data verification,  is as old as government.   In this case however it affects the fate of everyone in the world, every economy, the very earth itself.  The science really matters.

Adapting with Data

In conservation science its known as Adaptive Management and in concept is at the heart of the art of wildlife management.   As a young biologist in Georgia in the 70’s I worked with seasoned wildlife biologists, men who fought world war 2 and the Korean War , than followed it with graduate education through the GI bill. They worked in the time when game wardens fought Okefenokee Swamp-born poachers with guns of their own and not unaccustomed to using them.  As a recent yankee migrant,  the public meetings on new hunting regulations looked like a Star Wars bar scene.  But my bosses knew how to manage wildlife, and were smart enough  to develop early deer harvest models that mathematically linked deer age and condition in order to predict fawn production.  Using that, they would determine the following year’s deer harvest.

Deer and turkeys in a field

deer and turkeys in a field near Greenwich NJ. photo by L Niles

Today’s analytical capabilities extend this simple adaptive management system to new levels of complexity. It incorporates predictive modeling, complicated math that bewilders all but a few.  The models are so complicated, and the statisticians so certain of their scientific underpinning, they bristle at contrary observational data.   If the models predict 60,000 birds and biologists see 20,000 than the birds must be hiding!  Most biologists, managers, policy makers and citizens know little of these complex models and so cautiously rely on biometricians to sort it out for them.

So it is with the Adaptive Resource Management model being used by the Atlantic States Marine Fish Commission to determine the number of horseshoe crabs killed each year for bait.  Developed by a team of scientists from US Geological Survey, the US Fish and Wildlife Service and several state agencies (this author assisted in the development of the model).  It is a marvel of statistics.  It takes life history data for horseshoe crabs and red knots,  incorporates crab harvest data, current information on numbers for both species and models thousands of outcomes only to choose the most likely as a management recommendation.  If one were to listen to the scientists describing it  you would assume they be right.

Unfortunately the whole scientific edifice rests on the assumptions of the model.  The importance of assumptions came out spectacularly in the financial crash of 2007.  Predictive modeling made many people rich in the run up to the great recession.  Model-directed investments and hedges on those investments drew in money like flies on honey.  The math whizzes of Wall Street helped traders short, put and call the future for everyone.  What could go wrong?   For one thing the financial models assumed the catastrophic failures of the banking system of 1929 were a relic of those times and choose not to include this nasty bit of economics into the modeling data.  A lot of people lost their life savings on that bad assumption.

As with the economics models, there are many assumptions in the horseshoe crab model.  For example it only applies harvest restrictions to the known wintering distribution of Delaware bay crabs on the Atlantic Coast.  After breeding on Delaware Bay beaches, the crabs leave the bay in fall to winter on the ocean floor from NJ to Virginia.  This according to the king of horseshoe crabs Carl Shuster, for whom the ASMFC named an ocean reserve at the mouth of Delaware Bay.  Eschewing Shuster’s map the ASMFC’s choose new mapping that suspiciously allowed Maryland to harvest hundreds of thousand of crabs not included in the model’s predicted harvest.  When asked where else could they come, they had no answer. When I asked “Why not take them directly from Maryland’s small horseshoe crab population in Chicoteaque Bay”?   In response Maryland’s veteran marine biologist scoffed “it would wipe them out in a year”.


range map of Delaware bay horseshoe crabs

Winter range of Delaware Bay and Chesapeake Bay Horseshoe crabs. Notice the range include all of Marylands who disingenuously maintains only 50% of their 350,000 crab kill breeds in Delaware Bay


So what if Shuster is right and those crabs are from  Delaware Bay, what happens to the model?  It would be like following a car GPS system that is off by a few miles.  You will definite get somewhere, but not where you intend to go. Perhaps its no accident horseshoe crabs fail to respond to the enlightened model’s prognostications.

Smart Adaptation

Thankfully adaptive management doesn’t belong to statisticians.   The Paris Agreement illustrates how good adaptive modeling can be used by policy makers to define actions for the future.   Politicians, economists, financial markets, the public all want to know that whatever we do, it is fair and is working.  The statisticians don’t drive the system, they only help all the parties know what is going on and how it is changing.

Unfortunately for people and wildlife, adaptive management has fallen out of favor with many wildlife biologists.  Most work to solve thorny problems confronting wildlife.  Many ( including this author) focus on the promise created by advances in technology to discover new and exciting aspects of population or community ecology. A large portion would rather question scientific theory than the success of a habitat restoration, or the efficacy of a new regulation.  Rare is the case when biologists track the success of new wildlife conservation policy.  I know of none that actually spend time assessing the actions wrought by politicians on behalf of corporations – a troubling scenario for even the most optimistic among us.

For example, our Governor ( when he can squeeze in the time to govern) sees eye to eye with local politicians on the negative impact of conservation regulation to the economy.   In contrast, conservationists see habitat destruction everywhere and  argue for more..  Maybe both are right.  No one has really evaluated either the negative or positive impact of these regulations.  They just opine about it endlessly.

If one took the time to look at the data the picture is rather clear.  The time series mapping constructed by Rick Lathrop’s lab at Rutgers uses spectral imaging from LandSat satellites to form a compelling case that decades of very expensive land use regulation have not even slowed the onslaught of habitat destruction.  In short, we are wasting millions on a system that achieves neither developer’s or conservationist’s goals.

gif of land use change in NJ

land use change in NJ from the 70’s to 2000’s

But you see, one simple change, embedding an assessment of progress could have changed all that.  If we impose regulations and see they don’t work but cost plenty, than we can work to choose another way.   Thus starts the path to solutions.   Each new step informs the previous step and ultimately we can solve the thorniest of problems.