Saving the Planet
Planetary resource surveillance using geospatial analytics offers an innovative method for busting the bad guys
Human stewardship over the Earth’s resources is a weighty responsibility. We are by far the most numerous, widespread, and intelligent large mammals on the planet—as well as the sneakiest.
Humans are exploiting every known resource in every way imaginable. Try as we may to minimize depletion, pollution, and other activities that rob future generations of their inheritance, someone is always trying to evade the laws, regulations, policing, and other constraints we put in their way.
In the sphere of civil governance, surveillance has a bad reputation. However, when the surveillance promotes sustainability initiatives, we need to soften our cynicism. For example, no responsible person would recommend turning a blind eye as poachers exterminate endangered species. Likewise, the world needs the governments in the equatorial zone to maintain constant vigilance over rain forest destruction, which can exacerbate global warming and lead to a permanent loss of biodiversity.
Geospatial surveillance challenges
Geospatial analytics can be an invaluable tool for planetary resource surveillance. Ideally, that surveillance should be continuous, detailed, multidimensional, and global. It should draw on data from satellite imagery; ground-level sensors; public, private, and nonprofit organizations; crowdsharing communities; and other sources. And it should drive real-time response loops designed to pinpoint perpetrators and practices so they can be interdicted and sanctioned before environmental catastrophes show their ugly heads.
However, 24/7 geospatial surveillance of global resource utilization can be tricky. For starters, there’s the logistical challenge of monitoring a wide range of diverse activities, including mining, farming, forestry, fishing, manufacturing, and so on. Also, there is the challenge of continuously measuring all the relevant resource impacts of these activities. Furthermore, there are many actions that are infeasible or impractical to monitor to the extent that we would wish.
If resource-consuming activities take place on private property, there is a legal limit on how far the general public or even government regulatory authorities can snoop. But even if it occurs on public lands or waterways, illegal exploitation might be concealed because it takes place underground, or underneath thick forest foliage, or in a manner that makes it, at first glance, indistinguishable from legal utilization. And, of course, the perpetrators will usually falsify any self-reporting they might be required to produce to cover their tracks.
Ideally, geospatial analytics should find these hidden utilization patterns in whatever data is available, based on machine learning, statistical inferences, and predictive modeling. What’s needed are machine-learning algorithms and rich visualizations that can spot patterns consistent with poaching, overcutting, overfishing, erosive agricultural practices, and the like.
Zeroing-in activity detection
The recent article, “A Method for Predicting Fishing Activity Based on Geospatial Motion Behaviors,” presents a fascinating use of geospatial analytics that could help detect an illegal-fishing signal from the noise of too much data. Data scientists used geospatial positioning data and motion analytics to identify vessel behaviors on the open seas that were indicative of fishing in circumscribed waters. Data that fed their geospatial analytics included time codes, vessel identities, navigational status, rates of turn, speed over ground, latitude and longitude readings, true headings, and true bearings.
The specific patterns they searched for were circular and duplicative motions consistent with fishermen zeroing in on hitherto-undetected fish in open waters. “We noticed,” says the article’s authors, “that frequent and significant changes in the vessels’ compass heading (erratic heading) and erratic changes in velocity were strong predictors of fishing activity. The vessels themselves use a navigational status of 7 to self-report fishing activity, but this was under- and over-reported throughout the data set.”
Conceivably, the same techniques might also be used to detect zeroing-in activities associated with land-based resource depletion activities. For example, professional poachers tend to use vehicular mounts to scout and shoot their prey, which they may detect only after several circuits around the relevant habitat.
What these and other exploitation activities have in common is reliance on exploratory, trial-and-error methods to identify otherwise hidden resources. Multi-pass behaviors—circular and duplicative motions—are a dynamic pattern that may best be spotted through machine learning.
Please share any thoughts or questions in the comments.
|[followbutton username='IBMdatamag' count='false' lang='en' theme='light']|