Big Data lets you see the forest and the trees

Big Data lets you see the forest and the trees

Venus-C via Microsoft

The mainstreaming of Big Data analytics is rewriting the rules of forestry management. Tree farmers, plantations and other groups managing forestry assets can dig up relevant insights in a matter of hours using software on desktop computers.

The forestry industry has long used predictive modeling to forecast the environmental impact of planned harvests, controlled burns and other timberland management strategies. Traditionally, however, this sort of analysis has required dozens of individual, custom-developed spreadsheets, access to academic supercomputers and weeks of processing time.

This opportunity is capturing the attention of high-profile companies including Google and Microsoft, both of which are creating cloud-based forestry resources, but the most recognized name in forest analytics software is Canadian company Remsoft, in Fredericton, New Brunswick. 

"The software has greatly accelerated the insights that we need to make a new investments or be involved in competitive situations," said Timothy Cooney, director of research and analysis for Global Forest Partners, an investment company with more than $3.1 billion in timber assets under management across places such as Cambodia, Australia and New Zealand. "We can find the best answer, and we can answer our questions much faster."

It used to take Cooney's team of five researchers up to a year to develop sophisticated spreadsheet models, which were often specific to a certain region or forest. Remsoft's technology enables them to explore different management scenarios far more easily. "You can adjust your planning to fit the environment you're in," Cooney said.

Remsoft's technology, developed by two experts in forestry management, has actually been around for more than 20 years. Over that time the software has evolved to accommodate an increasingly larger number of variables, including the social impact of certain timber management strategies and environmental implications – everything from potential water stress to whether or not a specific plan might affect an endangered species.

"Usually there is some sort of risk that is forcing businesses to get more sophisticated with their analysis," said Remsoft's co-founder and co-CEO, Andrea Feunekes.

Analysis along three dimensions

Remsoft's software – which is used by investment companies, timber farmers, universities and government agencies – analyzes forestry management along three dimensions. It accounts for biological factors, such as how long a specific tree species takes to grow, the sort of habitat it requires and how much space it needs; along with spatial considerations, such as where certain harvest blocks are placed and when they can be cut. The software also supports the ability to assess the impact of decisions over long periods of time, not just in the moment, Feunekes said.

"It's a complex jigsaw puzzle," she said.

Aside from the productivity benefits cited by users such as Global Forest Partners, the return on investment (ROI) realized by some Remsoft customers has stretched into seven figures. For example, consulting and real estate company American Forest Management has used the software to help some of its properties improve the rate of return on their harvesting practices by up to 35 basis points, a big deal for an industry that boasts notoriously thin margins.

Elsewhere, the software is being used by Brazil's largest pulp and paper company, Suzano Papel E Celulose, to increase projection more sustainably.

"Since we implemented Remsoft's solutions, we can optimize assets that were once managed in various spreadsheets," the company said in a case study about the project. "As a result, we have greater alignment and communication between divisions and our productivity gains have reached 5 [percent] to 10 percent."

Forest analytics attracts heavyweights

Cloud computing technology also promises to reshape the resources that forest managers will have at their disposal. Remsoft, for one, is evaluating whether to provide access to its modeling capabilities within a cloud service. Meanwhile, Google and Microsoft have become involved in the development of cloud resources that address deforestation and wildfire management.

Google is involved in the development of the World Resource Institute's Global Forest Watch 2.0. The interactive, real-time forest monitoring system set to launch in late 2013 will use satellite technology, data collected with remote sensors, and human observation to aid forest management efforts. 

The tool will be available to commodities buyers and suppliers to better inform ongoing procurement decisions.  

“Deforestation rates in the Brazilian Amazon have dropped by 80 percent since 2004,” said Nigel Sizer, director of the Global Forests Initiative for the World Resources Institute, when the tool was introduced in April 2013. “This is in part due to their efforts to improve the quality and availability of information about what is happening to those forests and to make it rapidly available to those who can take action.”

The Microsoft cloud, meanwhile, is being used in Greece to host a new wildfire prevention application that can help predict daily wildfire risk on the island of Lesvos. There's an enormous amount of data behind the scenes, including weather patterns, topography considerations, vegetation and past fires. The application was developed on VENUS-C, a cloud resource that Microsoft has made available to research groups and startups. The application combines the capabilities of Bing Maps, Microsoft Silverlight and Windows Azure. It was created by the Geography of Natural Disasters Laboratory at the University of the Aegean in Greece.

“With the cloud computing infrastructure, we were able to do business as we couldn’t do in the past,” said Dr. Kostas Kalabokidis, associate professor, University of the Aegean, in a published case study about the project. "[Windows Azure] is essential for us, because the cloud provides us with the necessary processing power and storage that is required. That means the real end users for the fire department do not need to have any huge processing power or storage capabilities locally.