Issue link: https://read.dmtmag.com/i/75042
decisions regarding the province's natural resources," adds Geordie Robere-McGugan, inventory development specialist, Ontario Ministry of Natural Resources. In addition, a digital imagery-derived surface model and a five-meter classified digital surface model (DSM) were created. North West Geomatics and Leica Geosystems developed a DSM-generation tool for ADS data based on Semi-Global Matching, a new technolo- gy that creates a detailed elevation dataset referenced exactly to the imagery. The field-sampling component is designed to increase the interpreter's knowledge of the forest through mea- surement of forest-inventory ground plots. All field sampling is conducted according to the Ontario Forest Resources Inventory Calibration Plot Specifications. FRI calibration plots are 200 meters in length, with 10 stations located 20 meters apart. It's important that all plots are completed exactly on the ground as located on the aerial imagery. The information collected for each plot includes geographic location and general terrain information, a main stand- description summary, tree tally, age (see Figure 2), height, diameter, basal-area calculations, soils, ecosite, and stand profile. The sampling intensity is one plot per five square kilometers in forest-management units (FMUs), and one plot per eight square kilometers in other areas. Softcopy Interpretation The project's next component is image interpreta- tion. To carry out interpretation, most companies use softcopy systems (see Figure 3), which combine interpretation and data-automation phases. This improves the integration of supplementary informa- tion, supports interactive quality control and allows for efficiency in editing. All imagery classification is completed accord- ing to the Ontario Forest Resources Inventory Photo Interpretation Specifications. The information described for each forest polygon includes area classi- fication, source of information, productive forest modi- fier, management consideration, horizontal structure, incidental species indicator, stage of development, ecosite, area, vertical structure, species composition, leading species, age, height, crown closure, site index and site class. "Moving to a digital operation greatly enhances productivity," notes Radford. "Instead of needing to manually draw forest stands and then digitize them, the interpreter can directly digitize and capture each stand boundary and its attributes in one step. That will enable us to really streamline our workflow and reduce inventory production time significantly." Softcopy interpretation also allows interpreters to choose their preferred image format; use additional tools available to measure tree heights, slope, dis- tances and elevation; manipulate digital imagery to enhance ground-vegetation or tree-species identifica- tion; map polygons as small as 25 square meters; and integrate other sources of information such as geological surveys, disturbance maps, harvest and silvicultural information, and previous inventories. The interpreters must pass a mandatory ecological FRI training course as well as a species-identification test, establish a minimum of 1 percent of the forest- inventory calibration plots, obtain calibration-plot and supplemental data for the area, perform a supervised water classification using digital imagery, delineate and classify homogeneous polygons according to specifications, and submit to the OMNR for audit to confirm compliance with the specifications. To date, approximately 103 individuals have passed the image- interpretation test, and some 250 individuals have been trained for field sampling. After the FRI The inventory's fourth component will be a post-FRI assessment to establish approximately one plot per 20 square kilometers. These plots will be randomly selected based on a stratification of forest conditions within FMUs. The assessment plots will be established in the year Figure 3. The use of softcopy workstations enables the display of multiple information themes in stereo, enhancing the interpreter's ability to delineate and describe forest stands. 24 GEO W ORLD / JU L Y 2O12 after the forest-inventory production has been com- pleted, and will enable the development of accuracy and confidence estimates of inventory attributes. The data will provide for a statistically reliable analysis of forest-inventory attributes (e.g., age, height, species, ecosite) and will provide an opportunity to adjust the inventory for forest modeling. In addition, to monitor changes in forest condition, a network of permanent forest-inventory plots is being established. The plots will be systematically located Resource Management