This site hasn't been updated in a while, and there's a reason for that.  Two actually.  The first reason is that my wife and I had a wonderful spring visitor, a little baby girl, and she's taken up a lot of our free time this year. And while I am a very proud dad, I will spare her posterity and not post much more about that here.

The second reason I have not updated this site is that I have been busy getting ready for my new job in January, being an Assistant Professor at the University of Maryland, Baltimore County.  I am really looking forward to teaching and doing research in the Department of Geography and Environmental Systems at UMBC!  It's a great group of folks that I will be working with, and my first two courses are coming up this spring: Advanced Spatial Data Analysis, and Geographic Information Systems.  

Between continuing my postdoctoral research and starting new projects in Costa Rica and Maryland, I will have plenty of fun work for my new lab.  Eventually, this site will transition to be the lab web site, and other folks will make their appearances here as we work together. Things will get more formal, I guess, though I will probably blog more often.  

I will leave you will my favorite visualization of this past year, courtesy of the NYT and Japan's new weather satellite:
http://www.nytimes.com/interactive/2015/07/10/science/An-Image-of-Earth-Every-Ten-Minutes.html

 The Global Landscape Forum (GLF) finished in Lima, Peru today.  I am excited about their continued development of the integrated landscapes approach.  Supported by a wide variety of governmental and non-governmental organizations, the GLF seeks to break down the silos that exist among land-use specialists (fun Q: Why don't foresters talk to agricultural economists or water managers very often?  A: They go to different meetings and publish in different journals.).  

The integrated landscapes approach recognizes that multiple objectives and perspectives can be encompassed in the same landscape.  It argues that landscapes need to be managed as wholes, with certain land-uses better suited for different regions.  With adequate planning that conserves and connects different areas and recognizes the needs of different stakeholders, there is room for achieving many different objectives in any given landscape.  We can protect nature, feed the planet, enrich the poor, and sequester carbon to combat global warming and ocean acidification.  

It may sound like a pipe dream: have your cake and eat it too.  But it is important to remember that nature rarely requires ALL of a landscape to maintain biodiversity.  Similarly, people rarely require ALL of a landscape to produce food.  Between those two extremes, there is a real and undeniable opportunity to create the conditions where nature and humans coexist in a synergistic relationship.  One can image a landscape where hill forests protect water and lowland farms produce food.  Careful protection of selected lowland forests and river corridors could maintain many lowland species, and careful siting of mines and logging projects in the mountains could maintain regional water quality.  

And sure, there are plenty of cases where ecological realities, poor planning, or unfavorable economic conditions force win-lose choices in landscape planning.  A landscapes approach is not an exercise in blind optimism; it is simply recognizes that win-win situations can be promoted if stakeholders and experts work together.  An integrated landscapes approach often leads to a restoration roadmap for the way back from a win-lose choice.    One conservation-oriented success story is chronicled in the book "Green Phoenix", which details the restoration and expansion of a national conservation area in the dry northwest of Costa Rica.  

What do we need to know about landscapes to support informed, integrated decisions about land-use management?  I would argue that we need to know how different land-uses interact across space, and what the basic requirements are for different land-use objectives.  If we want clean ground water in our well, we should minimize pesticides on our farm, but water quality is also affected by one's uphill neighbors.  It is only by working together in a democratic, egalitarian process can we ensure clean water for all farms.  But who is more important for groundwater quality: uphill farmers near streams, or uphill farmers on steep slopes?   That's where good science comes in.  

Taking an example from my work in Costa Rica, if we want wildlife to persist while producing cattle, pineapples and bananas from the same area, we need to know several things.  We need to know where forests need to be protected, how to best protect them, how to effectively connect the forests so organisms don't get stuck on forest "islands" and go extinct, and where to reforest to reconnect isolated forest patches.  And that's just to protect the natural side of things; planning to sustainably increase production of agricultural crops in a forest-dominated region is even more complicated.  Science is integral to supporting this political process.  So, scientists and landscape planners: work with people who are planning together to care for their home, and you could save a forest near you today.  Or you could plant a farm with life-saving corn.  If you plan well, you could do both.  

This week, researchers (Tan et al. 2014) demonstrated an automated image classification algorithm that did better than the person-supervised classification approach for classifying hyperspectral imagery.  What does this mean?  One day I may be out of a job!  And it's great!  Let me explain why.  

One of the chapters of my thesis (in review for publication) dealt with classifying different types of forest in a hyperspectral image.  I wanted to track the expansion of tree plantations that were being subsidized by the Costa Rican government.  And boy, Internet, was it a tough row to hoe.  Working with hyperspectral imagery is computer-intensive, tricky, and full of pitfalls and challenges. But it is AMAZING what you can pull out of the data--there's just so much of it!  Some of my image files were 50 GB in size, and with them I was able to classify tree plantations with sufficient (and satisfying) accuracy and make a great map.  

I will give you that Tan et al.'s new algorithm is most likely incredibly computer-intensive, and it worked with a limited (and quite distinct) set of classes.  Obviously I am not quite out of a job yet.  But computer processing speeds are only increasing, and the cloud is doing more and more processing for scientists.  It is entirely likely that one day, when I start a hyperspectral analysis, 95% of my work may already be done by the computer.  

And I am perfectly happy with that.  The sooner that we can relegate supervised classification (a time-intensive, iterative data collection process)  to a minor component of the overall image analysis process, the better.  If I could've skipped the hard work of delineating forest from nonforest using an algorithm and gone straight to delineating tree plantations, that would have been great.  And if algorithms improve and I could skip the process entirely, just input my classes and get a tree plantation map, even better!  Nothing is sweeter in science than asking a great question and getting the answer the very next day.  

But wait, you may say.  Doesn't that mean that you, an image analyst, would be out of a job?  Well, there's a few things you should consider.  First, I am interested in using satellite and aerial imagery for conservation and answering ecological questions.  So automated techniques would only speed up my work and let me test more hypotheses.  Second, there will always be technical challenges to overcome; working with images from space or a speeding plane or UAV is not straightforward, and you need experience and expertise to problem-solve.  Third, there will always be new frontiers in remote sensing that require expertise and algorithm testing.  Hyperspectral lidar, high-resolution thermal imagery, and hyper-temporal satellite imagery are frontiers that are just now opening. 

I get excited just thinking of the new, innovative analyses that we will be able to do with new automated algorithms and new sensors.  Remote sensing and GIS are in their early decades still, and the science and science applications are expanding rapidly.  I have a friend who works with hyperspectral imagery of a different kind, of people.  He is using it to do accurate, automated classification of skin blemishes, to improve cancer screening.  And that, I think, is a good thing.  

Neglect and lack of funds will cause the U.S. to be blind to impending storms in 2017; our satellite fleet needs more money.  In a recent article, the NYT focuses on program mismanagment and cost overruns, but glosses over the perennial fight that NOAA faces for money to keep our weather satellites flying.  We need more earth-observing satellites, internet!