09-10, 10:30–11:00 (America/Chicago), Grand C
Raster Tools easily scales across various hardware configurations and lowers the barriers for parallel processing for spatial, statistical, and machine learning procedures; further advancing researchers’ and practitioners’ abilities to transform data into useful information for decision making.
Big data and the knowledge we glean from it are fundamentally changing the way resource management decisions are made. The use of remotely sensed data, ever expanding computer technology, and enhanced processing techniques provide natural resource managers with depictions of ecosystems at unprecedented spatial and temporal resolutions. While these sources of information are currently being leveraged to inform decision making, the sheer amount of data currently being collected has outpaced our abilities to efficiently manipulate and use those data for decision making. Newer tools, algorithms, and approaches are needed to address processing limitations and provide new opportunities to embrace the volume, variety, and velocity of big data streams. Important questions related to data scale, relevance, transformation , as well as the types of tools needed to efficiently extract information for decision making are at the forefront of empirically driven decision making . To that end we have developed a python based geospatial processing library called Raster Tools that automates delayed reading and parallel processing. In this presentation, we highlight two case studies that use the newly developed Raster Tools package to perform spatial, statistical, and machine learning analyses.
The first case study uses Raster Tools, basal area ha-1, stem counts ha-1, and forest cover type raster surfaces to inform sample design and improve variable estimation for forest stand inventories within the Fort Stewart significant geographic area (SGA) located in Georgia, USA. While the raster surfaces provide useful information for prioritizing longleaf pine (Pinus palustris) restoration efforts, timber sale administrators will want to know the volume or weight of the merchantable timber within the stands identified for treatment prior to implementation. Estimating existing volume for various timber products across a forest can be expensive. To reduce the cost of volume estimation, timber sale administrators can use ancillary data that are highly correlated with volume data to substantially reduce sampling effort for a given level of estimation accuracy. To illustrate these concepts, we will present a Jupyter notebook that demonstrates how Raster Tools can be used to both identify restoration sites and summarize basal area data used to aid in planning a sample design that reduces inventory costs and estimation error.
The second case study uses Raster Tools and raster surfaces depicting basal area ha-1, stem counts ha-1, and spatially explicit error estimates to identify and prioritize longleaf gopher tortoise habitat restoration in Florida, USA. This example highlights the flexibility and efficiencies of Raster Tools, how Raster Tools can be integrated with conventional raster spatial modeling, and the use of fine grained spatially explicit information to aid in habitat restoration. Three key components to this example are 1) spatially defining existing and potential open pine gopher tortoise habitat, 2) evaluating the use of prescribed fire within existing and potential gopher tortoise habitat, and 3) integrating recreational aspects into the prioritization process. Like the previous use case, we will present a Jupyter notebook that demonstrates the utility of Raster Tools in prioritizing gopher tortoise habitat restoration.
At the forefront of this accelerated pace and scale of data driven decision making is the development of spatial, statistical, and machine learning techniques that fully leverage existing hardware and adopt newer processing strategies to integrate big data sources seamlessly and easily with the decision-making process. Packages such as Raster Tools facilitate this integration while also providing functionality that can be used to further our understanding of natural resources while simultaneously providing the computational framework to optimize and justify management decisions at both scale and extent. While these tools facilitate Big Data analytics, they also necessitate a broader understanding of the role of spatial data and analyses within decision making. Moreover, they highlight the need for easy access to and integration with various open source and proprietary software systems.
Raster Tools provides a great deal of flexibility and is available through Python’s package index. These tools are readily available, free, and provide the fundamental architecture to ingest big data streams directly into spatial, statistical, and machine learning analyses. Moreover, these tools are easy to use and can facilitate data driven decision making at fine scale across broad extents.