Home

PEM Predictive Ecosystem Mapping User's Guide

1. S4 3 4 1 5 Critical Issues for Using PEM in Timber Supply Analysis S4 5 4 2 Wildlife Habitat Interpretations S4 7 4 2 1 INTO ee ein S4 7 4 2 2 Overview of Wildlife Habitat Assessment S4 7 42 3 Modeling HGDW A laica iaa S4 8 4 2 4 Using PEM in Habitat Models and Assessments S4 11 4 2 5 Critical Issues for Using PEM in Habitat Analysis S4 11 March 28 2005 PEM Predictive Ecosystem Mapping Page iv User s Guide 4 3 Biodiversity Interpretations S4 12 Hol Modu HON os 54 12 230528077 niet 54 13 4 3 3 Using PEM in Managing for Biodiversity Objectives 54 13 4 3 4 Critical Issues for Using PEM to Help Manage for BiodiVersity ODJEELVES nn iii 54 15 4 4 Rare and Special Ecosystem Inventories 54 16 4 5 Riparian Management Areas 54 18 4 6 Archaeological Overview Assessment S4 20 4 7 Spatially Distributed Process Models 54 22 5 Data Types and Attributes S5 1 5 1 Spatial Inv
2. A7 1 March 28 2005 PEM Predictive Ecosystem Mapping Page v User s Guide List of Figures Figure 2 1 Milestones in the development of ecosystem classification and mapping in British Columbia S2 5 Figure 3 1 Raster data model U U S3 2 Figure 3 2 Vector or polygon data model S3 3 Figure 3 3 The process for undertaking a PEM inventory S3 3 Figure 3 4 Knowledge base iterative process S3 9 Figure 4 3 Projecting yield for timber supply analysis S4 5 Figure 4 4 Wildlife habitat capability for wintering ungulates S4 10 Figure 4 5 Using PEM to identify potential range areas in NDT 4 S4 14 Figure 4 6 Using PEM to delineate areas that may support rare or special ecosystems 54 17 Figure 4 6 Using PEM to delineate riparian ecosystems 54 19 Figure 4 8 Using PEM to delineate areas that may have supported food gathering by indigenous peoples 54 21 Figure 5 1 Big BEC in the Cariboo Forest Region S5 2 Figure 5 2 TRIM CONQUE NAD RE S5 3 Figure 5 3 Bioterrain map polygons S5 4 Figure 5 4 Targeted terrain map
3. Figure 2 1 Milestones in the development of ecosystem classification and mapping in British Columbia 1950 1960 1970 1980 1990 2000 2010 Site classification concepts BC biogeoclimatic ecosystem classification Forest regions of Canada BC adopts BEC produces regional guides Soils and terrain mapping p Biophysical habitat mapping Terrestrial ecosystem mapping Correlation of data BEC refinement Predictive ecosystem modeling TEM standards EcoGen o EcoGen RIC certification standards for ecosystem inventories data collection PEM standards eo oe PEM accuracy protocol 00 it We thank David Moon for helping sort out the distinction between TEM and PEM March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S2 6 What Is PEM Ecosystem Classification The biogeoclimatic ecosystem classification BEC system developed based on European approaches to vegetation ecology integrates vegetation site and climatic parameters BEC has been modified through the years and the B C Ministry of Forests MOF adopted the BEC system as the basis for forest management in 1976 Over the subsequent decade each forest region developed a standardized ecosystem classification system based on BEC and produced regional Field Guides to Site Identification and Interpretation Terrain Mapping Soils mapping conducted by the federal department of agricultu
4. A similar problem applies in a raster environment Each cell derived from the terrain polygon will have proportions of relevant terrain assigned to it making the raster PEM result aspatial within the cell The ability to predict exact locations of site series based on terrain will be no better than with vector Depending upon the heterogeneity of the targeted terrain polygons expert knowledge must be used to include variability within the targeted terrain polygon and subsequent variability within the site series prediction Errors in variables such as species composition crown closure stand height stand structure all result in errors in the PEM prediction where those variables discriminate between site series or structural stages As above Site series predictions based on bedrock differences may be in error Errors in location of terrain attributes will produce errors in the site series location prediction Soils maps will have the same limitations as terrain and bioterrain maps Inadequate training and testing of the imagery classification will lead to errors in the PEM prediction Even if spatially accurate it is unlikely that field data will register exactly with other input data layers Errors in plot or transect UTM will result in erroneous assessment of the PEM result when plot data are compared with PEM output Errors in field attribute classification will result in an erroneous assessment of PEM input layer attri
5. See for example DeLong C 2003 A field guide for site identification and interpretation for the southeast portion of the Prince George Forest Region B C Min For Res Br Victoria B C Land Manage Handb No 51 10 D Moon 2005 Personal communication 11 Jenny H 1941 Factors of soil formation a system of quantitative pedology McGraw Hill New York 1 Major J 1951 A functional factorial approach to plant ecology Ecol 32 392 412 1 Demarchi D A E C Lea M A Fenger and A P Harcombe 1990 Biophysical habitat mapping methodology first draft B C Ministry of Environment Wildlife Branch Victoria B C March 28 2005 PEM By early 2005 about 15 and 53 of the province had been mapped with TEM and PEM respectively Much of this mapping has been accepted in the provincial digital data warehouse managed by the Ministry of Sustainable Resource Management Predictive Ecosystem Mapping User s Guide Page S2 7 What Is PEM landscape features of terrain at scales of 1 20 000 and 1 50 000 This classification was used to develop wildlife habitat interpretations The mapping of BEC site series evolved from biophysical habitat mapping using the MOF standard nomenclature for ecosystems and the TEM system TEM approaches and standards were adopted and the Resources Inventory Committee RIC developed the first certification for ecosystem inventories and ecological data collection in 1998
6. PEM TEM Alternatives http www for gov bc ca hre temalt TEM amp PEM Home Page http srmwww gov bc ca ecolo tem index html Problem Analyses of the TEM Alternatives Task Force http www for gov bc ca hre temalt public htm Standards http srmwww gov bc ca ecolo tem manuals html TEM http srmwww gov bc ca rib wis tem VRI Ecosystem and Habitat Mapping http srmwww eov bc ca rib wis vriehm March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A5 1 Appendix 5 TEM Projects in BC Appendix 5 TEM projects in British Columbia 1 50 000 TEM Project Areas 1 20 000 TEM Project Areas Bioterrain Project Areas Source Province of B C March 28 2005 Predictive Ecosystem Mapping User s Guide Page A6 1 Appendix 6 PEM Projects in British Columbia ses z Appendix 6 PEM projects in British Columbia 1233 PEM Project Areas Source Province of B C March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A7 1 Appendix 7 Appendix 7 Frequently Asked Questions What is the difference between vector and raster based PEM models The rules used to convert real geographical variation into discrete objects points lines polygons is referred to as a data model a set of guidelines for organizing the data in a database All PEM models are based on either raster grid or vector polygon data models or a combination
7. Assuming that FORECAST predictions can be made for individual or groups of site series PEM maps can be used to help scale up the predictions from the individual locations or sites for which the detailed models have been run to other locations or sites with similar attributes as associated with PEM ecological classes A second potential role for PEM maps in process modeling is to provide some of the input data required to parameterize these models Most process models require a very large amount of information about a large number of site factors or conditions One of the main challenges is to assign realistic and credible estimates of likely properties conditions or states to each and every physical location for which the model is used to make a forecast There is potential for consulting PEM maps and other similar types of maps to help prepare estimates of properties e g porosity depth hydraulic conductivity texture litter thickness and conditions e g moisture content depth to water table temperature for which a model equation is applied PEM maps cannot supply all inputs required to run physically based process models but they can help improve estimates of many inputs There is real potential for PEM maps and various ancillary data layers prepared in the construction of PEM maps to be designed so as to maximize their utility as sources of input data for physically based modeling 1 Kimmins J P Mailly D and Seely B
8. Ecosystem Modeling As computing improved and became more cost effective ecological mapping models were developed in western Canada as a low cost surrogate for air photo interpretation based methods of ecosystem inventory The Ecological Land Data Acquisition Resource system ELDAR a polygon based method of spatial inventory was developed by the Alberta Research Council and first tested in British Columbia in the McGregor Model Forest Prince George in 1994 The B C government supported the development of another polygon based model as a low cost alternative to TEM in the former Prince Rupert Forest Region in 1996 This model formed the basis of the present day EcoGen PEM model During this time remote sensing researchers throughout North America and Europe were also developing ecological mapping models By 2001 various PEM models and projects were in progress throughout British Columbia and RIC developed standards for mapping output and documentation The first standards were published for publicly funded PEM projects in 1998 with subsequent standards released in 2000 and errata in 2004 A protocol to test the accuracy of mapping using independent data was developed in 2000 and 2003 by the MOF 4 By early 2005 about 15 and 53 of the province had been mapped with TEM and PEM respectively see Appendices 5 6 2 2 What Is a PEM Inventory A PEM inventory includes four output products or components 1 site series polygon layer
9. PEM Predictive Ecosystem Mapping User s Guide Page A2 2 Appendix 1 Glossary Fuzzy logic A deductive system predicated on the notion that truth is a multi valued continuous quantity It allows for the concept of partial truth truth values between completely true and completely false It might be more correctly called continuous logic Geo referenced data located by a system of coordinates usually by GPS can be in the Universal Transverse Mercator UTM or longitude and latitude systems Knowledge base in the context of PEM the formal expression of the rules used to predict site series from ecological attributes the knowledge base comprises both the set of data tables mapping the relationships between ecological attributes and site series and the algorithms used to interpret the data and relationships to predict site series Landscape facet a distinct feature or element of landscape Low site in timber supply analysis in British Columbia a category of site productivity applied to stands considered not productive enough for timber harvesting and management Map entity is the basic ecosystem element being mapped e g Ecosection Biogeoclimatic Zone subzone variant phase or site series structural stage modifier seral community type It can be simple or complex Map feature a point line or polygon representing a site linear feature or area on a map Map features may be labelled with multiple map
10. 2 site series attribute database which contains for each polygon BEC variant 14 Meidinger D 2003 Protocol for accuracy assessment of ecosystem maps B C Ministry of Forests Research Branch Victoria B C March 28 2005 PEM The shape of the PEM output polygon can be based on site series result landscape shape bioterrain or forest Cover Predictive Ecosystem Mapping User s Guide Page S2 8 What Is PEM deciles for site series modifiers for slope and aspect optional 3 structural stage polygon layer 4 structural stage database PEM and TEM produce essentially the same outputs but they differ in how the outputs are produced and documented and the consequent cost of production 2 2 1 Comparing PEM to other inventories Several types of ecological inventories are used in British Columbia Table 2 1 Biophysical habitat bioterrain and TEM polygons are based on permanent landscape shapes formed from the deposition of materials by glaciation water wind and gravity and subdivided by directional exposure and BEC variant Forest cover polygons are based on stand structure species composition age and height at the time of air photo interpretation they are ephemeral shapes reflecting a combination of the tree species and stand structure at the time of air photo interpretation the history of the stand and the growing potential of trees Vegetation Resource Inventory VRI polygons can be base
11. March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S2 5 What Is PEM reliance of some PEM models on soils mapping or bioterrain mapping see below clouds this distinction At one time a method that relied on expert knowledge to classify or predict outcomes would have been called an expert system but this term seems to have fallen out of favour in recent years In the PEM type of expert system combinations of input data e g site characteristics such as BEC variant slope and aspect are assigned to specific interpretations e g site series by an ecologist with local experience and or using statistical relationships derived from plots The rules of interpretation relating ecological characteristics to ecosystem classification are formally structured in the knowledge base Terrain mapping delineates and classifies surficial materials landforms and geomorphological processes Bioterrain mapping utilizes the principles of terrain mapping as well as delineation of aspect and soil moisture classes in the surficial materials landforms and geomorphological processes Biophysical habitat mapping is a precursor to TEM in which polygons are delineated from bioterrain mapping and ecosystems are identified within the bioterrain polygons 2 1 3 Origins and Evolution of PEM The classification and mapping of British Columbia s diverse ecosystems represent a significant effort over the past 50 years Figure 2 1
12. documented which facilitates its use in a variety of habitat supply models 4 2 5 Critical Issues for Using PEM in Habitat Analysis The following critical issues should be addressed when making PEM based interpretations for habitat analysis Accuracy and Reliability of the PEM model The PEM model output should have some stated measure of accuracy that can be related to the required reliability of the interpretations Errors in the PEM model will be carried through the habitat analysis Appropriateness for the species of interest The habitat requirements of the species need to be based on attributes of the PEM model If the model does not provide the level of detail required to predict the habitat attributes then it is inappropriate for the interpretation For example PEM output is not appropriate for the prediction of habitat use by carnivores unless it is somehow tied to the habitat requirements of the prey species Suitability of PEM polygons for use in the habitat model If the PEM produces large complex polygons it may be difficult to make meaningful habitat interpretations about entire polygons How should one interpret critical habitat that is only proportionally March 28 2005 PEM PEM mapping can be used to assist in establishing biodiversity emphasis options through the identification of the BEC variant site series and structural stage relative to the natural disturbance type NDT P
13. informally structured rules of interpretation PEM usually relies on computerized application of formally structured rules of interpretation the knowledge base Predictive Ecosystem Mapping User s Guide What Is PEM Page S2 4 2 1 2 Ecosystem Mapping Ecosystem mapping is the stratification of a landscape into map units according to a combination of ecological features primarily climate physiography surficial material bedrock geology soil and vegetation Common scales of ecosystem mapping are 1 20 000 to 1 50 000 though larger scales such as 1 10 000 or 1 5000 may be used depending on project objectives Ecosystem maps along with associated interpretations supply valuable information for many applications related to the Forest and Range Practices Act including landscape unit planning Timber Supply Review and range use planning These maps also can be used to develop and apply the Biodiversity guidelines the Riparian Management Area guidelines and the Identified Wildlife Management Strategy Terrestrial Ecosystem Mapping TEM is a method of ecosystem mapping that relies on manual as opposed to automated application of relatively informal rules of interpretation and is largely based on air photo interpretation and field checking It is often but not necessarily used at large scales where detailed information is required The usual TEM mapping unit is the site series level of the biogeoclimatic ecosystem clas
14. modeling Predictive Ecosystem Mapping User s Guide Page S3 1 How PEM Works 3 How PEM Works 3 1 Assumptions PEM is based on three key assumptions 1 The combination of climate topography or landscape surficial geology soils water vegetation animals and microorganisms and the relationships among them give rise to predictable mappable entities called ecosystems 2 Relationships among the elements of nature that make up ecosystems can be generalized in an ecosystem classification system such as the BEC system 3 Ecosystem maps can be interpreted in secondary analyses that can be used for natural resource management applications 3 2 Raster and Vector Types of PEM Models The variety of PEM models and types of spatial analysis used to produce PEM output reflects both the ingenuity of PEM practitioners and the differences in context and input data found across the province British Columbia has standardized its requirements for ecological classification database spatial data and meta data used in PEM projects funded by provincial government agencies and initiatives Regardless of the model used to generate ecosystem predictions all publicly funded PEM models must be converted to the provincial standard vector format for storage in the provincial data warehouse All PEM models are based on either raster grid or vector polygon data models or a combination of both However discussions of the relative merits of
15. special ecosystem interpretations riparian area management and archaeological overview assessments 4 1 Timber Supply Analysis This section discusses timber supply analysis and the use of estimates of site productivity in timber supply analysis the ways in which a PEM inventory can be used to estimate site productivity critical issues that should be considered when contemplating using a PEM inventory as input to timber supply analysis 4 1 1 Introduction The supply of timber is an important factor in planning and managing natural resources It is a key component of the information considered by the provincial chief forester when determining allowable annual cut AAC Timber supply analysis is the procedure used to estimate timber supply and to show how it changes with changes in assumptions about policies management practices and inventory data Table 4 1 It also supports land use planning processes such as development of Land and Resource Management Plans LRMPs and forest management planning e g silviculture strategies wildlife habitat management strategies 4 1 2 Overview of Timber Supply Analysis Timber supply the rate at which timber is made available for harvesting is a measure of the potential flow of logs out of the forest in cubic metres per year Timber supply is not the same as the inventory of timber which is measured in cubic metres The inventory of timber represents the maximum amount that cou
16. A EI A IS 7 J ISU ER T IUD P de a i s p x ra LET n NR COR URS a PIS PT duca RIS ES N A E Whe e SD MT PIENS N Ei N EX a gt C e fi is Y W FA ES PEN ve Y 2i Ut 5 y A a ae ae IE np A PA VIA NT Source JMJ Holdings Inc and Okanagan Innovative Forest Practices Society March 28 2005 PEM The final output of a PEM model gives the user only the BEC variant site series and modifiers for slope and aspect Predictive Ecosystem Mapping User s Guide Page S5 12 Data Types and Attributes Figure 5 11 In raster based PEM polygon boundaries are based on raster cells with the same predicted site series Em a 77 al see Z 2 d E A Sy AN SS e Source LandMapper Environmental Solutions and Weldwood of Canada 5 3 2 Map Entities Layer The format of the output of a PEM model is determined by client needs as revealed in the requirements analysis and in the case of publicly funded projects the PEM standards in place at the outset of the project Generally PEM model output data base contains the following fields e ECPtag which links the map entities database to the spatial polygon coverage FCODE which identifies the feature code of the polygon linework e SRC Fcode the source of the feature Sdecl which identifies the decile proportion of the most common site series in the polygon Site SI the site series number of
17. Although the methods used to relate the spatial input variables to the site series or map unit vary with the model all PEM models use a knowledge base define that relationship 3 4 1 Derivation of Knowledge Base Knowledge bases can be derived in different ways depending upon the PEM methodology The knowledge base scoring can simply be the product of expert ecological knowledge Vegetation ecologists with a sound understanding of the ecology of the BEC subzones or variants within the project area arbitrarily determine the weighting of each spatial input variable to site series based on their understanding experience or beliefs about the relationship between the site series classification and the input spatial data If the project area lacks extensive spatially accurate field data where site series have been identified then the expert knowledge approach can be used successfully and cost effectively Accuracy assessments of the Canim Lake and Quesnel PEM projects indicate that the use of a landscape approach based on expert knowledge of site series and March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 8 How PEM Works landscape relationships can be equal to or superior to knowledge bases trained with field data If the project area has been field sampled providing spatially accurate ecologically reliable data describing site series as part of the PEM project or from other initiatives within the p
18. B C Ministry of Small Business Tourism and Culture 2000 British Columbia Archaeological Inventory Guidelines Version 1 0 Archaeology Branch for the Cultural Task Force of the Resources Inventory Committee British Columbia Specifications and Guidelines for Geomatics Content Series Volume 3 Digital Baseline Mapping at 1 20 000 Release 2 0 January 1992 The new specification for TRIM II 1 20 000 and 1 10 000 Revision Data Capture Version 2 0 May 15 1997 Carmean W H 1975 Forest site quality evaluation in the United States Adv Agron 27 209 69 Clutter J L J C Fortson L V Pienaar G H Brister and R L Bailey 1983 Timber management a quantitative approach John Wiley amp Sons Toronto Ont DeLong C 2003 A field guide for site identification and interpretation for the southeast portion of the Prince George Forest Region B C Min For Res Br Victoria B C Land Manage Handb No 51 Demarchi D A E C Lea M A Fenger and A P Harcombe 1990 Biophysical habitat mapping methodology first draft B C Min Environ Wildlife Branch Victoria B C Eng M and D Meidinger 1999 A method for large scale biogeoclimatic mapping in British Columbia Version 1 0 B C Ministry of Forests Victoria B C Available from http www for gov bc ca hre becweb subsite map Bigbec pdf March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A3 2 Appendix 3 References Hills
19. ECPtag 82L032 7858 map sheet polygon number FCode WI25200100 feature code for polygon SRC FCODE feature code assigned to digitally copied arcs to identify the source Bgc vrt IDFdm2 BEC variant Sdec 1 8 80 decile dominant structure in the polygon Strct S1 7 structure code for the dominant structure in the polygon old forest Sdec 2 2 20 decile secondary structure in the polygon Strct S2 3 structure code for the secondary structure in the polygon shrub dominated 5 4 PEM Digital Data Standards The output and documentation of a PEM model funded by government has to meet the PEM standards in place at the initiation of the PEM project Standards have evolved with the PEM process over the past decade new standards are being proposed for implementation by 2006 Information on the PEM standards is available from the PEM website http www for gov bc ca hre temalt index htm 47 Resources Inventory Committee 1999 The standards for Predictive Ecosystem Mapping in British Columbia inventory standard Victoria B C March 28 2005 PEM CE On completing this section users will understand e the implications of input data quality to the quality of the final PEM output e examples of appropriate and inappropriate applications for PEM Predictive Ecosystem Mapping User s Guide Page S6 1 Limitations of PEM 6 Limitations of PEM 6 1 Data Quality Data quality refers to 1 attributes of data that i
20. G A 1952 The classification and evaluation of site for forestry Ontario Department of Lands and Forests Resource Division Report 24 Hills G A and G Pierpoint 1960 Forest site evaluation in Ontario Ont Dep Lands and Forests Toronto Ont Res Rep 42 66 p Jenny H 1941 Factors of soil formation a system of quantitative pedology McGraw Hill New York Jones K R Ellis R Holt B MacArthur and G Utzig 2001 A strategy for habitat supply modeling for B C Draft Volume 1 Prepared for Habitat Supply Modeling Steering Committee Jones R K et al 1983 Field guide to forest ecosystem classification for the clay belt site region 3e Ontario Min Nat Resour 123 p Ketcheson M T Dool and and C Littlewood 2002 Development of procedures for mapping special ecosystems in the Arrow Timber Supply Area Unpublished report to B C Ministry of Forests Research Branch Victoria B C Kimmins J P Mailly D and Seely B 1999 Modelling forest ecosystem net primary production the hybrid simulation approach used in FORECAST Ecol Modelling 122 195 224 Krajina V J 1959 Biogeoclimatic zones in British Columbia Univ B C Dep Bot Vancouver B C Bot Ser No 1 1965 Biogeoclimatic zones in British Columbia Ecol Western N Amer 1 1 17 1969 Ecology of forest trees in British Columbia Ecol West N Am 2 1 1 146 1972 Ecosystem perspectives of forestry H R MacMillan Forestry Lectur
21. It is important that the client and PEM practitioner consider the interpretations to be made from the structure model before determining the classification They also must consider the limitations of the input data sources used for the structure model The structure codes used in the structural stage layer are shown in Table 5 3 Table 5 3 Structure codes used in the structural stage layer Code Structural stage description 1 sparse bryoid less than 10 vegetation cover 2 grass forb dominated 3 shrub dominated 3a low shrub dominated 2 m 3b tall shrub dominated 2 to 10 m 4 pole sapling dominated 5 young forest dominated 6 mature forest dominated 7 old forest The final structural stage polygon can be derived in the same way as a site series polygon through a knowledge base and spatial input data see Section 5 3 1 The result can be a simple or complex polygon reflecting a single structural classification per polygon or up to three March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S5 14 Data Types and Attributes structural classifications per polygon The structural stage classification can be modified to suit the requirements of subsequent interpretations The structural stage database is linked to the spatial polygon coverage by the ECP tag Table 5 4 shows an example record for the structural stage layer Table 5 4 Structural stage layer for IDFdm2 8723 Field Example Description Entry
22. along the y axis All neatline boundaries contained the appropriate labels and attributes NA None 5 3 Output Data Format 5 3 1 Polygon Source and Format The source of the final PEM polygon output shape can be derived from existing polygons such as forest cover Figure 5 9 bioterrain Figure 5 10 or subdivisions of them or the final PEM polygon shape can be derived from the output of the PEM model itself Figure 5 11 The nature of the final polygons should be determined at the outset of the PEM project The interpretations drawn from the PEM output should consider the source and format of the final PEM polygon shape The final PEM output consists of a spatial coverage of polygons with a database describing the contents of the polygon Depending on the PEM model the polygons can be simple or complex small or large Minimum polygon size is generally set at 1 ha Usually no more than three site series are described within a polygon Complex polygons do not spatially depict the location of individual site series simple polygons do March 28 2005 Predictive Ecosystem Mapping User s Guide Page S5 11 Data Types and Attributes Figure 5 9 Forest cover based PEM o Predicted Site ean a A A x an Series Classes BEE EHEN N Forested Units 02 11 o MM 12 o EB 13 oT 05 Non Forested 06 Units 07 08 A Source Province of B C Figure 5 10 PEM polygon boundaries can be based on bioterrain TA
23. be subjected to the same planning regimes For example in NDT 4 there is a requirement that a portion of the landscape be maintained in good condition as cattle range PEM can be used to identify such candidate range areas Figure 4 5 The PEM structure model can also be used to identify areas of older structural stages that need to be considered in planning These units can be combined with other synthesis products from PEM TEM or mapping that depicts ungulate winter range critical grizzly habitat or other critical habitats This information may be combined with other spatial planning tools to depict the areas constrained for operational forestry Maintaining Rangeland The IDFdm2 BEC Variant is NDT 4 characterized by historically frequent and widespread stand maintaining fires One biodiversity objective in this NDT 4 variant is to maintain at least 85 of it as permanent rangeland Site series 02 Antelope brush Bluebunch wheatgrass could be targeted as an ecosystem appropriate for range area retention by using PEM output March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S4 14 PEM Applications Figure 4 5 Using PEM to identify potential range areas in NDT 4 BEC variant boundary I BEC variant decile site series letter code site series number aspect modifier IDFdm2 8 AW 02 w 2 RO 00 w e 80 Antelope brush bluebunch wheatgrass 20 Rock outcrops warm aspect Serub b
24. be used to predict the location of evidence of past use by First Nations It is only useful for pre stratifying the area of interest to assist in designing the inventory 50 http wlapwww gov bc ca habitat fish_protection_act riparian documents regulatio n pdf 51 B C Ministry of Small Business Tourism and Culture 2000 British Columbia Archaeological Inventory Guidelines Version 1 0 Archeology Branch for the Cultural Task Force of the Resources Inventory Committee March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A 1 j Appendices Appendices Appendix 1 Acronyms Appendix 2 Glossary Appendix 3 References Appendix 4 Information Sources Appendix 5 TEM Projects in BC Appendix 6 PEM Projects in BC Appendix 7 Frequently Asked Questions March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A1 1 Appendix 1 Acronyms Appendix 1 Acronyms AAC allowable annual cut BEC biogeoclimatic ecological classification DEM Digital Elevation Model ELDAR Ecological Land Data Acquisition Resource system FIA Forest Investment Account FSSIM Forest Service Simulator GWM general wildlife measures IWMS Identified Wildlife Management Strategy LRMP Land and Resource Management Planning MOF Ministry of Forests NDT natural disturbance type OGSI Old Growth Site Index PEM Predictive ecosystem mapping RIC Resources Inventory Committee RISC Resources Information Standards Committee RM
25. eere eene nennen S5 5 Figure 5 5 Landscape profile viii S5 6 Figure 5 6 Landscape faceta rai S5 6 Figure 5 7 Soil moisture prediction S5 7 Figure 5 8 Satellite imagery example S5 9 Figure 5 10 Forest cover based PEM example S5 11 Figure 5 11 Bioterrain based PEM example S5 11 Figure 5 12 Raster based PEM example S5 12 List of Tables Table 2 1 Comparing ecological inventories used in British Columbia S2 9 Table 3 2 Knowledge base example truncated S3 7 Table 3 3 Criteria for evaluating PEM for a specific application S3 10 Table 3 4 Considerations when specifying PEM attributes S3 10 Table 4 1 Steps in timber supply analysis to support AAC determiliatiQIi essen namen S4 2 Table 4 2 Site index estimates derived from PEM site series and SIBEC are used in four steps of the timber supply analysis process nun ann nun nn nn nun nun nn nn S4 4 Table 4 3 Direct and indirect effects of yield tables on timber supply S4 4 Table 4 4 Ecosystem structure based wildlife habitat ratings S4 9 Table 5 1 Input data quality assessment example S5 9 Table 6 1 Limitations of spatial data
26. entities Map unit see map entity Meta data data describing data meta data describe the content quality condition and other characteristics of the data of interest Precision used here in the non statistical sense of exactness or fineness of resolution with which a feature or property is described While in some degree related the statistical definition refers to the standard error or repeatability of measurement not the exactness or resolution of the measurement Some statistical textbooks use the terms precision and reliability as synonyms which is not the usage here Moon 1999 Predictive ecosystem mapping a computer GIS and knowledge based method of stratifying the landscape into ecological map units typically site series Rare ecosystems ecosystems labelled as red or blue listed by the Conservation Data Centre of the Ministry of Sustainable Resource Management March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A2 3 Appendix 1 Glossary Reliability thematic and spatial accuracy and precision one component of data quality Simple map entity a map entity representing a single site series Site index a measure of potential site productivity the capacity of an area of land to grow trees of a given species A simple definition is the height of dominant trees at age 50 Site series BEC all sites capable of producing the same mature or climax plant communities within a biog
27. for use in PEM S6 1 Table 6 2 Appropriate and inappropriate uses of PEM data S6 3 March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S1 1 Introduction 1 Introduction 1 1 Audience The intended audience for the Predictive Ecosystem Mapping PEM User s Guide is professionals who are planning to use an existing or planned PEM inventory for a particular application such as wildlife habitat interpretations or estimates of site productivity The audience includes resource management planners timber supply analysts systems ecologists and habitat specialists The guide is intended to stand alone as well as provide background reading material for the complementary workshop It is not a technical reference for designing or carrying out a PEM inventory 1 2 User s Guide Organization This guide is organized around describing PEM and its uses limitations data requirements and applications in British Columbia Each section begins with an indication of learning objectives 1 2 1 Content Section 2 describes the evolution of ecosystem mapping in British Columbia and explains the range of inventories that can be included in the term predictive ecosystem mapping Section 3 describes the types of PEM models a framework for undertaking a PEM inventory and critical assumptions It also addresses the issue of suitability of an existing PEM inventory for a specifi
28. of both In a raster data model the mapped area is divided into a regular grid of cells Every location in the mapped area corresponds to a cell in the raster Discrete objects are represented as single cells or groups of related cells In a vector data model the location of a discrete point in the mapped area is identified with reference to map co ordinates x y Discrete objects are represented as points single locations lines groups of related points or polygons groups of related lines A separate database contains the descriptive attributes of the objects with an index linking each object to a record in the database Data can be converted from raster to vector and vice versa so the choice of data model is an issue for the modeler since it affects the ways in which the data are used but not for the client who is really interested in applying the results What is a map entity A map entity is the thing that is mapped which in PEM is the site series What s the difference between a simple polygon and a complex polygon A simple polygon contains only a single site series or map entity A complex polygon can have up to three site series or map entities that are allocated proportions of the polygon e g 50 3576 15 Is bioterrain mapping a required element of a PEM model Bioterrain mapping or any type of terrain mapping may be used in a PEM model but is not required What is Big BEC and why is it important w
29. practices specified in strategic or landscape level plans WHAs conserve those habitats considered most limiting to a given species Generally identified wildlife provisions do not address habitat supply habitat connectivity and population viability and other issues such as access management Such issues should be taken into account during strategic or landscape level planning Species requiring consideration within strategic level plans are typically wide ranging species that are sensitive to landscape level changes such as badger caribou grizzly bear Marbled Murrelet Queen Charlotte Goshawk Spotted Owl and wolverine The IWMS species accounts describe the habitat of species to the level of Broad Ecosystem Unit which are amalgamations of different groups of site series and may include many distinct plant associations Grizzly Bear and Caribou Habitat Habitat use by grizzly bears and caribou is part of an information set that can be incorporated into spatially explicit forest harvesting models The attributes of those habitats can be wholly or partially described at the BEC variant site series and structural stage 4 2 3 Modeling Habitat Wildlife Habitat Ratings The modeling of wildlife habitat includes a variety of approaches some of which draw upon ecological mapping One form of wildlife habitat modeling is the wildlife habitat ratings approach used by the Ministry of Sustainable Resource Management 4 The ratings are based
30. results and synthesizes the results of the sensitivity analyses March 28 2005 PEM PEM is used to predict site series which are used in turn to estimate site productivity a key input to timber supply analysis Predictive Ecosystem Mapping User s Guide Page S4 3 PEM Applications 4 1 4 Using PEM in Timber Supply Analysis The main purpose of PEM inventories is to predict site series In timber supply analysis the site series predictions are in turn used to estimate site productivity Site productivity is usually represented in timber supply analysis with site index which affects or is used in the first four steps of timber supply analysis Table 4 2 Site Index Site index is used to represent site productivity based on the assumption that tree height accurately indicates growth potential 2 Many factors other than site productivity may reduce height growth confounding the reliability of tree height as an indicator of site productivity As a result site index estimates shown in forest inventory polygons are often inaccurate because they have been derived from measurements or estimates of stand or tree height and age This is especially true for very young and vary old stands Various approaches to characterizing site quality directly from site characteristics have been explored in the literature for many decades In recent decades much effort has been devoted in British Columbia to investigating and developing
31. the most common site series in the polygon Site series modifiers describing slope and aspect classes for the most common site series in the polygon The above are followed by Sdec 2 Site S2 and modifiers for the next most common site series in the polygon and by Sdec 3 SiteS 3 and modifiers for the least common of the three site series identified in the polygon if there are more than two Table 5 2 shows an example record for the map entities layer March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S5 13 Data Types and Attributes Table 5 2 Map entities layer for site series IDFdm2 6 DS 03w 4 DT 01 Field Example Description Entry ECPtag 82L032 7858 map sheet polygon number FCode WI25200100 feature code for polygon SRC FCODE feature code assigned to digitally copied arcs to identify the source Bgc vrt IDFdm2 BEC variant Sdec 1 6 60 decile dominant ecosystem Site S1 03 site series number code dominant ecosystem Site mcS1 DS site series letter code dominant ecosystem Site m1a W modifier w for warm Site m1b additional modifier if needed Sdec 2 4 40 decile second ecosystem Site S2 01 site series number code second ecosystem Site mcS2 DT site series letter code second ecosystem 5 3 3 Structural Stage Layer The structural stage layer of a PEM model predicts the structure of the stand using a classification approved by the regional ecologist and client at the outset of the PEM project
32. the proportions of site series predicted by the model compared with the proportions of site series found by the field sampling Then the knowledge base could be adjusted to produce a model result that when summarized predicts the same proportions of site series as do the field data This may result in a model that can reliably predict site series proportions but may not depict them accurately in the map Field data can be used to train the model by comparing the site series found in the field at the UTM coordinate of each plot to that of the model at that location This method must be used cautiously remembering that errors in input data and the knowledge base can result in the model 21 Moon D 2005 Personal communication March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 9 How PEM Works predicting site series in locations that are tens of metres away from their actual location Field data should be distance buffered or generalized to account for cumulative spatial errors The ecologist should examine field data relative to patterns of site series distribution on the ground The knowledge base then needs to be adjusted to produce realistic depictions that although perhaps not in the exact locations of the field data result in site series patterns that bear a strong similarity to those found in the field Figure 3 5 Knowledge base iterative development Model Building Process Spatial O input laye
33. value Precision is the exactness of measurements or predictions Precision generally decreases as the requirement for accuracy increases and vice versa The reliability required for any project is a function of the sensitivity of interpretive or predictive procedures and the consequences of errors If small changes in input values cause large responses in predicted values the requirement for precision is high Conversely if large changes in input values cause small responses in predicted values the requirement for precision is low If the consequences of predictive error are low the need for accuracy is low whereas if the consequences of error are high high accuracy is needed Some applications such as the Timber Supply Review require the PEM output to meet specified levels of accuracy determined by a third party using an independent data set Many ecosystem mapping projects are funded through the Forest Investment Account All FIA funded projects must be completed in accordance with the applicable provincial Resources Information Standards Committee standards The B C MOF Research Branch and Moon et al 23 have developed protocols for accuracy assessment of ecosystem maps There is interest in also making this level of accuracy a requirement for PEM to be used for wildlife habitat interpretation Not every PEM project has a required level of accuracy and not every PEM model developed meets all accuracy requirements Both TEM and PE
34. 1999 Modelling forest ecosystem net primary production the hybrid simulation approach used in FORECAST Ecol Modelling 122 195 224 March 28 2005 PEM CE On completing this section users will understand e essential and optional data sources e assessment of input data quality e the function and importance of the knowledge base in a PEM e format requirements and standards for PEM output The two critical spatial input layers for PEM are Big BEC and TRIM Predictive Ecosystem Mapping User s Guide Page S5 1 Data Types and Attributes 5 Data Types and Attributes PEM models use existing spatial inventory data to predict site series and structural stage It is critical to understand which data sources are essential to the model and which are optional All PEM models require the most up to date BEC lines and the TRIM planimetric base for map registration Additional sources of spatial data can be added to the model depending upon financial resources and the interpretive needs of the client as well as the quality and utility of the additional data to the model s predictive ability The sources of input data for PEM can be in various formats including vector raster and point This section overviews data sources commonly used in PEM models 5 1 Spatial Inventory Input Layers 5 1 1 Big BEC The biogeoclimatic ecosystem classification BEC is a hierarchical system that incorporates biotic veg
35. 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S2 3 What Is PEM species are sometimes used to ensure that the site unit has a unique name within the provincial classification Site series names use the same names as the site associations to which they belong preceded by the appropriate biogeoclimatic subzone or variant name or code Zonal site units are always numbered 01 Non zonal forested site series are numbered from 02 to 29 sequentially in order of driest to wettest moisture regime and secondarily in order of poorest to richest nutrient regime Site Classification Within any biogeoclimatic subzone or variant a recurring pattern of sites occurs reflecting the variety of site features and soil characteristics that occur across the landscape These ecosystems are described by the site classification of BEC Three units are formally recognized in the BEC site classification site association site series and site type Two other units site phase and site variation are not formally included in the classification but may be used to further subdivide site series Site association is the basic unit of site classification but site series is the unit most commonly used by operational field staff A site association includes all ecosystems capable of developing vegetation belonging to the same plant association or in some cases subassociation at the climax or near climax stage of vegetation development
36. A riparian management area SEI sensitive ecosystem inventory SIBEC Site Index Biogeoclimatic Ecological Classification TEM terrestrial ecosystem mapping THLB _ timber harvesting landbase TRIM Terrain Resource Information Management TSR Timber Supply Review VRI Vegetation Resource Inventory UTM Universal Transverse Mercator UWR ungulate winter range March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A2 1 Appendix 1 Glossary Appendix 2 Glossary Accuracy the closeness of the presented value to the true value Analysis unit in timber supply analysis the aggregate of hectares which have the same species composition and site index and will be managed in the same way i e hectares which are assigned the same yield tables Area based with respect to timber supply analysis models used to indicate that the relative location of each stand or cutblock topology is represented in the model Also called spatially explicit Big BEC revision of Provincial BEC lines to a larger scale typically 1 20 000 or 1 50 000 suitable for use in PEM Biophysical using landscape shape or terrain mapping to predict the occurrence of an ecological entity process or productivity Bioterrain mapping uses the techniques and principles of terrain mapping and also delineates aspect and soil moisture classes within the surficial materials landforms and geomorphological processes Capability The ability of the habitat
37. Consultants Inc and JMJ Holdings Inc developed the guide and accompanying workshop The project team is grateful for the support and direction received from the following individuals Project Manager Project Advisor User s Guide Reviewers Technical Contributions Nona Phillips David Moon Dave Clark Corey Erwin Earl Spielman David Moon Nona Phillips Tim Bogle David Carson Dave Clark Tim Harding Colleen Jones Bob MacDonald Bob MacMillan Del Meidinger David Moon Phil Winkle Nona Phillips Forestry Consulting CDT Core Decision Technologies Inc Ministry of Sustainable Resource Management Ministry of Sustainable Resource Management West Fraser Mills Ltd CDT Core Decision Technologies Inc Nona Phillips Forestry Consulting Ministry of Forests Timberline Forest Inventory Consultants Ltd Ministry of Sustainable Resource Management Tolko Forest Industries Ltd Ministry of Sustainable Resource Management Ministry of Sustainable Resource Management LandMapper Environmental Solutions Inc Ministry of Forests CDT Core Decision Technologies Inc Decision Tree Forestry Consulting March 28 2005 PEM Predictive Ecosystem Mapping Page iii User s Guide Table of Contents 1 I trodu uCtioiiik u IILI anna na S 1 UN PUOC APP E mST S1 1 1 2 Users Guide Organization scarico 1 1 1 2 1 E E e ai E E E E EE E S1 1 12 2 eique mr S1 2 oe aau us masa 51 2 2 What ISs IPEM7z u u obi GO ERU RER
38. In other words a site association is a group of related ecosystems physically and biologically similar enough that they have or would have similar vegetation at maturity Since a site association can include ecosystems from several climates it can be somewhat variable in its environmental conditions Therefore a site association is divided into site series within subzones and variants Although a site association occurs on ecologically equivalent sites the site series in it may occupy different positions on the scale of relative moisture and nutrients in different biogeoclimatic subzones or variants For example a site association that occupies sites that are relatively dry compared to others in a wet variant may be found in areas that are relatively wet compared to other sites in a dry variant Site series then are subdivisions of site associations and include all sites within a biogeoclimatic subzone or variant that are capable of producing the same mature or climax vegetation unit plant association Site series are described in the Regional Field Guides to Site Identification Site and soil conditions and the vegetation community are used to identify site series 3 http www for gov bc ca hre becweb publications becpubslist htm RFG March 28 2005 PEM Both TEM and PEM are methods of ecosystem mapping rely on expert knowledge and predict BEC variant and site series TEM relies mainly on manually applied
39. In the context of this guidebook the process of estimating site index involves two basic steps predict site series with PEM and then look up the site index in the Site Index Site Unit Report tables Assuming that a PEM prediction of site series is available for each forest polygon we can proceed to using the Site Index Site Unit Report tables to estimate site index as follows Since site index differs by species for any site the site index species must be chosen for each stand The site index species is the tree species for which site index is estimated the dominant species is usually used With zone subzone variant site series and species identified the site index can be looked up in the tables of the Site Index Site Unit Report by Biogeoclimatic Unit If the polygon contains more than one site series average site index should be computed weighting by the amount of the stand area in each site series For more information on using SIBEC refer to http www for gov bc ca hre sibec SIBEC_ RDM Section 6 htm 4 1 5 Critical Issues for Using PEM in Timber Supply Analysis Timber Supply Review TSR has special requirements for using PEM because of the legislated responsibilities and obligations that accompany the determination of ACC A third party accuracy assessment is March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S4 6 PEM Applications particularly important in this context The MOF Forest Ana
40. Land Resource Research Institute Agriculture Canada CEF Ottawa Ont Agriculture Canada Misc Publ No 84 29 Olivotto G and D Meidinger 2001 Development of EcoYield a conceptual model for timber supply analysis using predictive ecosystem mapping and site index ecosystem relationships B C Min For Victoria B C EcoNote 01 1 Pojar J K Klinka and D V Meidinger 1987 Biogeoclimatic ecosystem classification in British Columbia For Ecol Manage 22 119 154 Resources Inventory Committee 1997 Standards for wildlife habitat capability and suitability ratings in British Columbia Terrestrial Ecosystems Task Force Ecosystems Working Group Wildlife Interpretations Subcommittee Victoria B C 1999 The standards for Predictive Ecosystem Mapping in British Columbia inventory standard Victoria B C Rowe J S 1972 Forest regions of Canada Canadian Forestry Service Department of the Environment Ottawa Ont Publication No 1300 1980 The common denominator in land classification in Canada an ecological approach to mapping Forest Chronicle 56 19 20 1984 Forest land classification limitations of the use of vegetation Proc of the symposium on forest land classification March 18 20 1984 pp 132 147 Madison Wisc March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A4 1 Appendix 4 Information Sources Appendix 4 Information Sources EcoGen http www for gov bc ca hre ecogen
41. M are used throughout the province by many different users for many different reasons Assessing the thematic accuracy of each individual mapping project is a key component in the determination of the appropriate interpretative uses of the data Any PEM or TEM intended for use in TSR timber supply applications must meet certain minimum map accuracy standards 2 Meidinger D 2003 Protocol for accuracy assessment of ecosystem maps B C Ministry of Forests Forest Science Program Victoria B C Technical Report 011 23 Moon D E D Dunlop and N Phillips 2004 A protocol for assessing thematic map accuracy using small area sampling Unpublished report to the Cariboo Site Productivity Assessment Working Group 24 Meidinger D 2003 Ecosystem mapping accuracy and timber supply applications http www for gov bc ca hre becweb pdf PEMaccuracystatement2003 pdf This March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 12 How PEM Works The MOF has also adopted a standard procedure for the accuracy assessment of ecosystem maps A standardized approach allows for the comparison of results between different projects and or areas of the province and gives users a quantitative sense of data quality The accuracy assessment protocol can be used as part of a quality assurance procedure and or as a component of quality control undertaken by a TEM or PEM practitioner Third party contractors assess the accur
42. O ao S2 1 2 PO eub t een S2 1 2 1 1 The Biogeoclimatic Ecosystem Classification System S2 1 2 1 Ecosystem Mapping anni S2 4 2 1 2 Origins and Evolution OF PEM as S2 5 2 2 What Is a PEM Inventory een S2 7 2 2 1 Comparing PEM to other inventories S2 8 2 2 2 Overview of PEM inventory applications S2 10 3 How PEM Works RR RE KR VD ICT EDD RU i S3 1 SS ME usse pte S3 1 3 2 Raster and Vector Types of PEM Models 53 1 3 3 Modeling Framework for PEM eee nmm nnn S3 3 3 4 Knowledge Base nes eds S3 6 3 4 1 Derivation of Knowledge Base S3 7 3 4 2 Testing Knowledge Base S3 8 3 5 Assessing the Suitability of a PEM Inventory S3 9 3 6 Accuracy and Reliability Assessment S3 11 4 PEM Applications in Resource Analysis S4 1 41 Timber Supply Analysis en een S4 1 41 1 Modu net naar AAA TT S4 1 4 1 2 Overview of Timber Supply Analysis S4 1 4 1 3 Modeling Timber SUDO au S4 2 4 1 4 Using PEM in Timber Supply Analysis
43. PEM data warehouse 3 4 Knowledge Base The knowledge base of a PEM model relates variables from the spatial input data Section 5 1 to the site series or map unit classification approved for use in the PEM model by the MOF Regional Ecologist The PEM practitioner and an ecologist with local experience build the knowledge base by assigning relative rankings to a matrix of site series and attributes This process captures the ecologist s knowledge of the occurrence of site series on the landscape and applies it to the data in a consistent manner Knowledge bases can be structured many different ways depending upon the individual PEM model A knowledge base may consist of rows and columns within a spreadsheet where each input variable or combination of input variables is assigned a value or weighting for a site series or group of site series Table 3 2 Some knowledge bases are simple use very few variables and have direct relationships between the variable and site series Other types of knowledge bases infer relative relationships between spatial variables and the site series through cumulative scores with the site series having the highest score being assigned to the spatial location of the appropriate combination of variables That location could be an individual raster cell or a pre determined polygon such as a bioterrain or forest cover polygon or a subdivision of those polygons Some practitioners have developed knowledge base
44. Predictive Ecosystem Mapping L 4 r us i f air cN wl March 2005 Predictive Ecosystem Mapping User s Guide March 28 2005 Prepared for Tolko Industries Ltd Prepared by Cortex Consultants Inc and JMJ Holdings Inc PEM Predictive Ecosystem Mapping Pagei User s Guide Foreword Predictive ecosystem mapping PEM is a type of expert system that has been applied to large scale classification and mapping of ecosystems across British Columbia It is seen by many as a useful alternative to the more costly terrestrial ecosystem mapping TEM This guide provides an overview of PEM and its appropriate application to potential users including resource management planners timber supply analysts systems ecologists and habitat specialists March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Acknowledgements Page ii The guidebook and associated workshop were prepared for Tolko Industries Ltd Williams Lake Woodlands Project funding was provided by the Land Base Investment Program LBIP of the Forest Investment Account a program of the government of British Columbia to support sustainable forest management practices improve the public forest asset base and promote greater returns from the utilization of public timber This project was conceived by Tyler Mitchell and John Stace Smith of Riverside Forest Products Cortex
45. ability of the PEM model appropriateness for the species of interest e suitability of the PEM polygons for use in the habitat model Predictive Ecosystem Mapping User s Guide Page S4 11 PEM Applications 4 2 4 Using PEM in Habitat Models and Assessments Spatial and database information generated by a PEM model can be used in habitat models and assessments In capability suitability models the BEC variant site series and structural stage information from the PEM database is related directly to the capability suitability ratings via the wildlife habitat ratings tables The result can be depicted as a map or used as a database Then the capability suitability information can be used to assist in the assessment of interpretations related to ungulate winter range or used as input in a habitat supply model The PEM ecosystem database and spatial coverage can also contribute directly to a wide variety of habitat supply models by providing an input layer much as PEM uses other spatial inventories as input Ecosystem can be one consideration in a habitat supply model the spatial relationships among ecosystems are another factor considered by some habitat supply models The effect of forest practices on habitat quality can be determined through a combination of ecosystem mapping from PEM and wildlife capability suitability look up tables in some habitat supply models PEM output as found in the PEM data warehouse is standardized and
46. abitat conditions for all native species Maintaining habitat diversity is used as a surrogate for maintaining biodiversity Nevertheless special efforts may be needed to protect the habitat of species known to be at risk such as threatened endangered or regionally important species Specific strategies for addressing these species are outlined in the Managing Identified Wildlife Guidebook The conservation of biodiversity depends on a coordinated strategy that includes a system of protected areas at the regional scale provision for a variety of habitats at the landscape scale and management practices that provide important ecosystem attributes at the stand scale PEM mapping can be used to assist in establishing biodiversity emphasis options through the identification of the BEC variant site series and structural stage relative to natural disturbance type NDT Within each NDT the present day seral stage distribution riparian habitats and 36 B C Ministry of Forests and B C Ministry of Environment 1995 Biodiversity guidebook Forest Practices Code of British Columbia Victoria B C Forest Practices Code guidebook March 28 2005 PEM PEM output can provide information on both the diversity of ecosystems found and the distribution of structure within those ecosystems at the time of mapping Predictive Ecosystem Mapping User s Guide Page S4 13 PEM Applications important wildlife habitats can be identified f
47. acy of Resources Information Standards Committee RISC standard TEM and PEM project deliverables The client is responsible for submitting via accuracy assessment contractors all applicable accuracy assessment documentation to the Province in accordance with the procedures outlined in the standards document document presents minimum map accuracy standards for the use of PEM and TEM in timber supply applications 25 Resources Inventory Committee 2000 Standard for Predictive Ecosystem Mapping digital data capture version 1 0 Available from http srmwww gov bc ca risc pubs teecolo pemcapture assets pem pdf March 28 2005 PEM CE On completing this section users will understand e the variety of potential interpretations or applications for PEM inventories e the ways in which PEM is used in different applications Timber supply analysis is a process of exploring how alternative forest management strategies and timber harvesting levels affect timber supply Predictive Ecosystem Mapping User s Guide Page S4 1 PEM Applications 4 PEM Applications in Resource Analysis A wide range of resource management activities could draw upon ecosystem mapping for input into decision making The two most commonly used applications for PEM are considered in detail in this section timber supply analysis and wildlife habitat assessment Other applications discussed include biodiversity interpretations rare and
48. areas that may have supported food gathering by indigenous peoples The Fd Snowberry Balsamroot site series is characterized by the presence of Balsamroot Balsamorhiza sagittata an important source of food to indigenous people BEC variant decile site series letter code site series number aspect modifier March 28 2005 PEM PEM maps may be useful for supporting the application of physically based or process models such as FORECAST by helping to scale up their predictions and by providing some of the input data required to parameterize these models Predictive Ecosystem Mapping User s Guide PEM Applications Page S4 22 4 7 Spatially Distributed Process Models Note that this discussion is somewhat speculative We thank Bob MacMillan of LandMapper Environmental Solutions for taking the time to make these observations PEM maps may be useful for supporting the application of physically based or process models such as FORECAST Kimmins et al 1999 FORECAST is an ecosystem based stand level forest growth simulator which uses relatively simple measures of decomposition nutrient cycling light competition and other ecosystem properties to simulate forest growth and ecosystem It uses local growth and yield data often from TASS TIPSY combined with other data to derive estimates of the rates of key ecosystem processes related to the productivity and resource requirements of selected species
49. butes and PEM site series result March 28 2005 PEM Table 6 2 PEM application Timber supply analysis Wildlife habitat prediction Biodiversity objectives Rare and Special Ecosystems Riparian Area Management Archaeological overview assessment Predictive Ecosystem Mapping User s Guide Limitations of PEM Page S6 3 6 2 Examples of Appropriate and Inappropriate PEM Data Applications The successful application of PEM for natural resource management depends on the suitability of the PEM data for the particular interpretation Although PEM data have many potential applications not all are appropriate Table 6 2 Appropriate use of PEM data Site index estimates Ungulate seasonal Capability Suitability Grizzly bear seasonal Capability Suitability Wetland dependent species Capability Suitability Grassland dependent species Thematic generalization of PEM BEC variants site series and structural stage output to drive seral stage distribution targets Stand structure interpretations should be made with caution the PEM stand structure model must be designed with biodiversity interpretations in mind PEM model should be designed with interpretation for rarity or special ness as part of the algorithm Rare and special ecosystems should be predicted with caution A few riparian area attributes required by the Riparian Area Regulations can be derived from PEM output These include ac
50. c application or conversely the design requirements for PEM intended to be used for a specific application Accuracy and reliability assessment are also addressed at the end of this section Section 4 discusses possible applications of PEM inventories such as timber supply analysis wildlife habitat interpretations biodiversity interpretations and rare and special ecosystem inventories Section 5 focuses on data types format and standards for PEM input and output data Section 6 discusses limitations of PEM related to input data and modeling methods March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S1 2 Introduction Appendices include acronyms and terminology technical references maps showing PEM projects across the province and frequently asked questions FAQs 1 2 2 Conventions Shaded notes to the reader are placed in the margin Shading within the text is used to highlight examples Italics indicate a term that is defined in the glossary Appendix 2 or legislation Bold text indicates emphasis 1 3 PEM Support Appendix 5 provides URLs for sources of current information on PEM in British Columbia March 28 2005 PEM G On completing this section users will understand e the basic elements of the BEC system the differences among ecosystem mapping Terrestrial Ecosystem Mapping TEM Predictive Ecosystem Mapping PEM terrain mapping bioterrain mapp
51. ccessional status In order to arrange ecosystems at the three levels of integration the BEC system combines four classifications vegetation climatic zonal site and seral Vegetation classification is most important to developing the ecosystem classification However the climatic and site classifications are the principal classifications used in the application of the BEC system At this time the seral classification has not been adequately developed Naming BEC units Biogeoclimatic zones are usually named after one or more of the dominant climax species in zonal ecosystems the Alpine Tundra Zone is a self explanatory exception and a geographic e g coastal interior or climatic modifier e g boreal montane Biogeoclimatic zone names are 1 Much of this material was extracted from the Ministry of Sustainable Resource Management TEM PEM Website http srmwww gov bc ca ecology tem index html 2 http www for gov bc ca hre becweb aboutbec aboutbec system basic htm March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S2 2 What Is PEM often referred to by a two to four letter acronym For example the Interior Cedar Hemlock Zone is referred to as the ICH Zone and the Montane Spruce Zone is referred to as the MS Zone Subzone names are derived from classes of relative precipitation and temperature or continentality The first part of the subzone name describes the relative precipitation a
52. ccurate information at 1 20 000 Spatial reliability questionable May be spatially suspect poor GPS data Other data limitations Overestimation of water in dry climatic areas Accuracy and reliability of photo interpretations Polygons may include small areas of other elements not noted in legend Accuracy and reliability of photo interpretations Data attributes may be modeled rather than photo interpreted Geologic units may be complexes that are not Spatially explicit Soils units may be complexes that are not Spatially explicit Imagery may yield same spectral quality for different substrates e g rock and bare soil Field classification of attributes may be erroneous Page S6 2 Consequences An indication of a stream within a gully when no stream exists increases the amount of wet habitat predicted by the model along the gully and at the bottom of the slope associated with the gully PEM using TRIM II will still have the same issues with micro topography as TRIM I Accuracy assessment may find errors based on BEC variant distribution that are not a consequence of the PEM as site series distribution along BEC boundaries may not reflect typic features If BEC line is not correct then the model s predictions in that area will also be in error Terrain variability is not spatial at the polygon level consequently the PEM output may not reflect terrain variability accurately within the polygon
53. ces of error are high high accuracy is needed Accuracy standards and protocols for accuracy assessments have been developed for using PEM in critical applications such as the Timber Supply Review See Section 3 6 of the User s Guide for more information Why doesn t my PEM give me a map PEM output consists of spatial files and a database If the client wants a printed map the spatial data and database have to be linked and queried to produce a plot file which depicts the elements of the database the client wants to see on a map Standard PEM output does not include map plot files If a requirement map production should be specified in the requirements analysis phase of the project Can the vector based standard output stored in the provincial data warehouse be subsequently rasterized Yes a polygon can be shattered back into grid format The data attached to each raster cell will retain the complexity of the original polygon If a complex polygon is rasterized each cell will also have the same group of site series as the original polygon Rasterization cannot render aspatial attribute data spatial Why does the PEM data base contain both a number and letter code for site series Early in the development of ecosystem mapping standards the BC Ministry of Environment MOE used two letter codes for ecosystems while the Ministry of Forests MOF used numbers When the two agencies cooperated to create mapping standards they retai
54. d on either stand attributes or a combination of stand attributes and permanent landscape shapes derived from bioterrain PEM models are based on the potential of the landscape to consistently give rise to a single site series or combination of site series The shape of the PEM output polygon can be based on site series result landscape shape bioterrain or forest cover March 28 2005 G00Z SZ YEN Sulojs sooa enads 6 75 oDed e e e e e e e e e e e e buiuuejd juubui jun suoneoijdde uellediy YSL ede spue bullaauibuy uorew ojul 103u9Aut Jo uoneoijddy oJnjonags jedse pue adojs e e JOJ PIJIPOUI salles DIS WIL SONIEA 9I1N0S91 19470 suqep Apoom 351809 uoneyoboaA e e Jaujo samads ean IMA o1njyinags AjAnonpoud e SUS SONDIAS DA ABAOD 1S 104 obeuieJp padse e UOneoyissep uie uieJ1o93o0lg oJnjonags jeygey e uwaj3s soda jes s Aydo g JUBWSSISSP Buiddew Buiddew Buiddew JEUGEY OJIIPIIM puezs 359104 Seues SYS ureJe9jolg 3343 94 d ayul odAj WOAJ paap edeus UOB JOd psyewony 0J0Ud 7U9 U0 INeWOUyL JOJU9AUT eiquinjo ysn11g ui pasn s riol1u Aul e9160 093 Burieduioo T Z S qeL ldd S JEUM opin5 s asn Buiddew ure3s so23 9AnpDIpeld vwd PEM h The focus of this guidebook is to explain how a PEM inventory works and how the PEM inventory could be used to support various applications Predictive Ecosystem Mapping User s Guide Page S2 10 What is PEM 2 2 2 Overvie
55. de enough information to be used to predict the location of these areas The reliability and scale of input data must be carefully considered before using PEM for this interpretation 40 B C Ministry of Forests and B C Ministry of Environment 1995 Riparian management area guidebook Forest Practices Code of British Columbia Victoria B C Forest Practices Code guidebook http www for gov bc ca tasb legsregs fpc fpcguide riparian March 28 2005 Predictive Ecosystem Mapping User s Guide Page S4 19 PEM PEM Applications Figure 4 7 Using PEM to delineate riparian ecosystems 70 of the target polygon supports a riparian ecosystem Spruce Horsetail BEC variant boundary IDFdm2 7 SH 07 k 3 SP 04 k er BEC variant decile site series letter code site series number aspect modifier IDFdm2 7 SH 07 k 3 SP 04 k tas E gt ar a ee 70 Swx Horsetail e 30 FdLw Spruce Pinegrass warm aspect Antelape hrush Scrub birch Bluehunch Horsetail wheatgrass 2 Colluvial R Redrock mixed Fluvial K Bekock milly cakar Lacintrine eus maleriali March 28 2005 PEM CE Exercise caution when using PEM for Archaeological Overview Assessments Human beings choices of habitat location are only somewhat predictable Predictive Ecosystem Mapping User s Guide Page S4 20 PEM Applications 4 6 Archaeological Overview Assessment PEM output can guide archaeologists to sit
56. dict site series using tree species presence of rock non productive brush alpine canopy closure etc It can also be used to predict stand structure Predictive Ecosystem Mapping User s Guide Page S5 8 Data Types and Attributes 5 1 6 Forest Cover or VRI Forest cover or Vegetation Resource Inventory VRI polygon data are often used in PEM models In some models PEM output polygon shape is based on forest cover polygons Forest cover polygons are based on tree species composition and site productivity Forest cover data can also be used as a spatial input data layer in PEM to assist in the prediction of site series where the presence or absence of a tree species or combination of tree species helps define the site series Other forest cover attributes can be used in a PEM model to aid in the prediction of site series such as the presence of rock non productive brush alpine wetlands open range and canopy closure Forest cover data can also be used in the structural stage portion of the PEM model Attributes commonly used to predict stand structure include age species composition height and history The format of VRI data is different than forest cover data as are the number of variables available to the PEM model The quality and detail of VRI data vary with the project although VRI data must meet standards 5 1 7 Satellite Imagery Satellite imagery which is in raster format has been used in PEM to help differentiat
57. e Series Univ B C Centre for Continuing Education Vancouver B C pp 1 11 _____ 1977 On the need for an ecosystem approach to forest land management In Ecological classification of forest land in Canada and Northwestern U S A Can Inst For For Ecol Working Group Univ B C Vancouver B C pp 1 11 Mah S and G D Nigh 2003 SIBEC site index estimates in support of forest management in British Columbia B C Min For Res Br Victoria B C Tech Rep 004 Major J 1951 A functional factorial approach to plant ecology Ecol 32 392 412 Meidinger D 2003 Protocol for accuracy assessment of ecosystem maps B C Ministry of Forests Research Branch Victoria B C March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A3 3 Appendix 3 References Meidinger D and J Pojar compilers and editors 1991 Ecosystems of British Columbia B C Min For Special Report Series No 6 330 p Moon D E 1999 Problem analysis on data quality assessment issues Draft Submitted to TEM Alternative Task Force by CDT Core Design Technologies Inc 2005 Presentation to Cariboo Site Productivity Working Group Moon D E D Dunlop and N Phillips 2004 A protocol for assessing thematic map accuracy using small area sampling Unpublished report to the Cariboo Site Productivity Assessment Working Group Moon D E and C J Selby 1989 Land resources inventory of the Power River Watershed
58. e physiognomic classes of vegetation Figure 5 8 It can assist in the delineation of brush from forest alpine tundra from krumholtz graminoid wetlands from shrubby wetlands and non vegetated areas like recent cutblocks rock snow and ice from vegetated areas The reliability of the satellite imagery needs to be determined before it is applied in the model March 28 2005 Each spatial inventory used in PEM must be assessed for input data quality Predictive Ecosystem Mapping User s Guide Page S5 9 Data Types and Attributes Figure 5 8 Satellite imagery Satellite imagery is used in the delineation of bare soil rock trees shrubs roads 5 2 Input Data Quality Assessment Each spatial inventory described in Section 5 1 and any additional spatial inventory used in PEM must be assessed for the quality of the data The input data quality assessment report a requirement of the PEM standards must be completed for each set of spatial data used in the PEM model The data must be reconciled to TRIM and positional accuracy reported relative to TRIM Many PEMs use retrofitted FC1 forest cover data which include reconciliation of the hydrographic features to the TRIM base Retrofitted maps will pass the input data quality protocol but do not permit evaluating the size of the original spatial errors or the spatial integrity of the lines that do not conform to the reconciled hydrography or control points The inpu
59. ea is identified with reference to map coordinates Discrete objects are represented as points single locations lines groups of related points or polygons groups of related lines A separate database contains the descriptive attributes of the objects with an index linking each object to a record in the database Both types of input data can be used in either data model raster can be converted to vector and vice versa The choice of data model is an issue for the model builder since it affects the way in which data are used it is of much less concern to the client who needs the site series predictions Figure 3 1 In a raster data model lines and polygons are generated from adjacent cells with the same attributes indicated by colour in this example III I I I cell I oe PT EE cells with same attributes March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 3 How PEM Works Figure 3 2 Ina vector or polygon data model objects are represented as points lines or groups of related lines TON o lo points e di e vectors olygon p un Q polygon 3 3 Modeling Framework for PEM The PEM framework comprises eight main activities Figure 3 3 Figure 3 3 The process for undertaking a PEM inventory 1 Requirements Analysis 2 Data Assembly Knowledge Base Development 1 Field Data GJ 5 PEM Classification 6 Internal Accuracy Assessment ee a pass pass 8 Final Rep
60. entory Input Layers S5 1 Sl TE BE o O EAE S5 1 Bb IRIS sS sss us sss S MEDI Kd S5 2 5 1 3 Tenan MAD DING susanne S5 4 5 1 4 Landscape Facet Modeling eese nene S5 6 5 1 5 Soil Moisture Models a rr S5 7 LO Forest COVER or VR ia aaa cito S5 8 541 7 S ji Image ornato S5 8 5 2 Input Data Quality Assessment eene nnne S5 9 5 3 OUDUE Das TO IC uuu u uu E i S5 10 5 3 1 Polygon Source and Format u S5 10 5 3 2 Map Entities Layer seiten een S5 12 5 3 3 Suelta Stage Lay OF een S5 13 5 4 PEM Digital Data Standards S5 14 6 Limitations of PEM iii eO GEORG GERA ERGO REC ERG RR dH DR S6 1 Sn Data OU rr G S6 1 6 2 Examples of Appropriate and Inappropriate PEM Data Applications S6 3 ADDENGICES na A1 Appendix 1 POT OMNY GINS zes seen Pu A1 1 Appendix 2 GIOS e A2 1 Appendix 3 oca ee mu lod Fo TREE EEG A3 1 Appendix 4 Information Sources A4 1 Appendix 5 TEM projects in British Columbia A5 1 Appendix 6 PEM projects in British Columbia A6 1 Appendix 7 PEM projects in British Columbia
61. eoclimatic subzone or variant Spatially explicit Not formally defined but usually used in the context of timber supply analysis to mean that the location of stands or cutblocks with respect to one another topology is represented in the model Special ecosystems ecosystems considered rare or unique uncommon interesting sensitive or of concern Stand structure see structural stage Structural stage the seral status of the site series based on a simple classification of overall stand structure one of bryoid grass forb shrub pole sapling young forest mature forest old forest Suitability the ability of the habitat in its current condition to support the species Timber harvesting land base in timber supply analysis the area considered available for timber harvesting alternatively the hectares on which timber harvesting is allowed in the timber supply model Timber supply the volume of timber made available for harvesting annually measured in units of cubic metres per hectare Timber supply analysis a process for estimating timber supply March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A2 4 Appendix 1 Glossary Triangulated irregular network TIN a form of digital elevation model based on irregularly spaced nodes in which each sample point has X Y co ordinates and a surface or Z value These points are connected by edges to form a set of non overlapping triangles which represe
62. eport to B C Ministry of Forests Research Branch Victoria B C 39 http srmwww gov bc ca sei March 28 2005 PEM Predictive Ecosystem Mapping User s Guide PEM Applications Page S4 17 Figure 4 6 Using PEM to delineate areas that may support rare or special ecosystems Castlegar and Surrounding Area Arrow TSA Rare Ecosystems PEM Field Plots Incorrect Plot Correct Plot Rarity Index mI No Rare Ecosysterns 1 Rare Ecosystem E 2 Rare Ecosystems 3 Rare Ecosystems E 4 Rare Ecosystems Source JMJ Holdings Inc March 28 2005 PEM CE Carefully consider the reliability and scale of input data before using PEM for RMA interpretations Predictive Ecosystem Mapping User s Guide Page S4 18 PEM Applications 4 5 Riparian Management Areas PEM can assist in the delineation of Riparian Management Areas RMAs The PEM output can identify features of the riparian area that can be used to assist in meeting the objectives of the RMA Figure 4 7 The outer edge of the active flood plain can be determined from the bioterrain input layer Wetlands are identified in PEM and depending upon the site series classification can be classified into simple and complex wetlands Riparian ecosystems used for grazing can be identified from PEM output The limitations to using PEM to assist in the delineation of RMAs are based in the resolution and accuracy of the PEM The input data may not provi
63. es that may have been important food sources for indigenous peoples and can assist in planning a sampling strategy for Archaeological Overview Assessments PEM could be used to assist in identifying areas that in the opinion of a trained archaeologist could have supported the living activities of early civilizations Such interpretations assist archaeologists in the pre stratification of the landscape into units with varying potential for evidence of utilization This process is a requirement for some types of forest development planning Site series 03 Fd Snowberry Balsamroot of the IDFdm2 BEC variant is characterized by the presence of Balsamroot Balsamorhiza sagittata This common plant provided an important source of food to indigenous people All parts of this plant are edible leaves can be eaten or smoked the taproots were roasted or steamed and dried and the seeds were pounded and used like flour Deer and elk graze on Balsamroot throughout the year About 6076 of the polygon shown in Figure 4 8 is predicted to support the growth of this culturally important plant Use PEM cautiously for this type of interpretation PEM is directed at predicting ecosystems based upon features of site climate and vegetation Human beings choices of habitat location are only somewhat predictable March 28 2005 aa PEM Predictive Ecosystem Mapping User s Guide Page S4 21 j PEM Applications J Figure 4 8 Using PEM to delineate
64. etation and abiotic climate and site attributes at the provincial regional and site levels BEC lines are available from the MOF and are continually being refined and updated The scale of provincial BEC mapping is generally 1 250 000 The lines are conceptual they usually need to be refined for use in mapping and predictive models generated at scales of 1 20 000 to reflect local climatic and site conditions The first step in the PEM process is to determine the suitability of the BEC lines and site series classifications based on their vintage PEM is often preceded by Big BEC mapping which refines the legacy BEC mapping with large scale 1 20 000 BEC lines Big BEC provides local climatic variation at an operational scale MOF regional ecologists determine whether the Big BEC lines are suitable for the PEM process Figure 5 1 shows the area in the Cariboo region mapped with Big BEC The biogeoclimatic subzone and variant lines mapped using the Big BEC process form the first level of stratification in a PEM model It is the initial spatial stratification within which the model makes its prediction of site series 42 Meidinger D and J Pojar compilers and editors 1991 Ecosystems of British Columbia B C Min For Special Report Series No 6 330 p 3 Eng M and D Meidinger 1999 A method for large scale biogeoclimatic mapping in British Columbia Version 1 0 B C Ministry of Forests Victoria B C Available from htt
65. f the PEM map is required Whether or not minimum accuracy levels are required to meet project objectives mapping accuracy directly affects the reliability of an interpretation The quality of an interpretation is a function of PEM accuracy sensitivity of the interpretation to PEM accuracy i e errors in the PEM and accuracy of the predictive procedure 18 Meidinger D 2003 Protocol for accuracy assessment of ecosystem maps B C Ministry of Forests Research Branch Victoria B C Technical Report 011 1 Moon D E D Dunlop and N Phillips 2004 A protocol for assessing thematic map accuracy using small area sampling Unpublished report to the Cariboo Site Productivity Assessment Working Group March 28 2005 PEM Knowledge bases can be structured many different ways depending upon the individual PEM model Predictive Ecosystem Mapping User s Guide Page S3 6 How PEM Works If an interpretation is not highly sensitive to errors in a PEM prediction a map that failed a PEM accuracy assessment might still be used to produce relatively accurate interpretations Conversely if the PEM map does not provide all of the information needed for an interpretation a highly accurate PEM map could cause large interpretation errors 8 Final Report Maps Data and Meta data The PEM process should be documented in a final project report and in the meta data of the deliverables submitted to the provincial
66. he following species mule and black tailed deer white tailed deer elk moose caribou bighorn sheep thinhorn sheep and mountain goat UWRs for these species have been established in many areas of the province The confirmation and establishment of UWRs and related objectives across the province to support the FRPA is ongoing Identified Wildlife Management Strategy IWMS Identified wildlife are species at risk and regionally important wildlife that the Minister of Water Land and Air Protection designates as requiring special management attention under the Forest and Range Practices legislation This includes endangered threatened or vulnerable species of vertebrates invertebrates plants and plant communities 32 http wlapwww gov bc ca wld identified iwms2004 html March 28 2005 PEM Typically wide ranging species that are sensitive to landscape level changes need to be considered within strategic or landscape level plans Predictive Ecosystem Mapping User s Guide Page S4 8 PEM Applications Regionally important wildlife include species that are considered important to a region of British Columbia rely on habitats that are not otherwise protected under FRPA and are vulnerable to forest and range impacts Identified wildlife are managed through the establishment of wildlife habitat areas WHAs objectives for WHAs and implementation of general wildlife measures GWMs or through other management
67. hen undertaking a PEM project Big BEC refers to revised BEC lines that better reflect local variation in climate at scales of 1 20 000 or 1 50 000 The distribution of local site series is better reflected in a PEM when the BEC lines have been updated using the Big BEC protocol March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A7 2 Appendix 7 Does PEM require field data for model building Field data may be collected for use in PEM model building activities but is not required See Section 3 4 2 of the User s Guide for information on how field data can be used to test the knowledge base What is the difference between accuracy and reliability Reliability which comprises accuracy and precision is the probability of a specific interpretation not being wrong it is relative and context dependent Accuracy is the closeness of a presented value to the true value Precision is the exactness of measurements or predictions The reliability required for any project is a function of the sensitivity of interpretive or predictive procedures and the consequences of errors If small changes in input values cause large responses in predicted values the requirement for precision is high Conversely if large changes in input values cause small responses in predicted values the requirement for precision is low If the consequences of predictive error are low the need for accuracy is low whereas if the consequen
68. ify the the yields of timber These classes which are referred to as analysis forest inventory units are usually based on leading tree species and site productivity e g spruce leading stands with site index greater than 18 m Step 4 Stand yield is projected for each management regime to be applied Project to each analysis unit These projections are represented as timber stand yields yield tables that show stand characteristics at different ages e g timber volume per hectare average stem diameter Step 5 Current management activities including those that enhance timber Identify production e g planting thinning and those that maintain or management enhance other forest values e g wildlife habitat visual quality are activities and identified Management requirements and the area to which they requirements apply are identified Step 6 In each decade simulated the model harvests available timber up to Model the the target volume and then grows the forest The area available timber supply for harvesting each decade is limited by management requirements base case such as forest cover constraints and minimum harvest ages Step 7 Sensitivity analysis is carried out to evaluate sources of uncertainty Sensitivity in the data and the management assumptions analysis Step 8 The TSR Timber Supply Analysis Report summarizes the information Synthesis of that was used in the analysis states the base case harvest forecast
69. ildlife habitat Recommended Required Both Simple Optional Optional assessment Biodiversity Recommended Optional Both Either Optional Optional objectives Riparian Recommended Optional Both Either Optional Advantageous Management Guidelines Sensitive or Recommended Optional Simple Simple Optional Advantageous special ecosystem inventories Archaeological Recommended Optional Both Either Optional Advantageous overview assessment map entity the thing that is mapped in PEM it is site series simple map entity represents a single site series complex map entity represents a complex of more than one site series which cannot be differentiated by the model simple polygon a polygon in which there is a single map entity complex polygon a polygon in which there is more than one map entity per polygon March 28 2005 PEM Reliability is the probability of a specific interpretation not being wrong e g a 100 accurate map can have low interpretive reliability if its content is inappropriate for the interpretation Predictive Ecosystem Mapping User s Guide Page S3 11 How PEM Works 3 6 Reliability and Accuracy Assessment To confidently use PEM for natural resource management applications its reliability and accuracy must be determined Reliability which comprises accuracy and precision is the probability of a specific interpretation not being wrong Accuracy is the closeness of a presented value to the true
70. ility makes it possible to explore the effects of treatment timing location and patterns of stand development on timber supply Such models are sometimes referred to as area based spatially explicit or geo referenced because they recognize the spatial relationships between treatment units e g cutblocks Strategic planning most often uses an aspatial model meaning that information in the database e g forest class is aggregated so that the locations of forest stands within the planning area are not tracked in the model Strategic planning explores the big picture implications of different management strategies without considering block by block how those strategies would play out on the ground it is assumed that they could be implemented satisfactorily Some models can incorporate forest cover constraints that approximate the effects of spatial management objectives such as adjacency rules The strategic plan gives direction to more specific plans such as tactical and operational plans in which such details are considered explicitly 30 Meidinger D 2003 Protocol for accuracy assessment of ecosystem maps B C Ministry of Forests Research Branch Victoria B C 31 http www jsthrower com gy prod March 28 2005 PEM Ecosystem mapping offers a consistent framework for looking at habitats of individual species or groups of species at both the coarse and fine filter levels Predictive Ecosystem Mapping Use
71. imber supply estimates are directly related to the area of the THLB and the yields expected for that area the yield projections and changes in them due to changes in site index directly affect the volume available for harvesting Yield tables also affect timber supply indirectly through their interaction with green up ages and minimum harvest ages used in timber supply models Table 4 3 Table 4 3 Direct and indirect effects of yield tables on timber supply Effects Mechanism by which timber supply is affected Direct Yield table volumes directly affect volume available for harvesting effects Indirect Age at which green up height is reached affects the green up or effects disturbance forest cover constraints used to represent visual quality and biodiversity requirements Age at which minimum harvestable volume per hectare or diameter is reached affects the minimum harvest age constraints which determine when young stands first become available for harvesting March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S4 5 PEM Applications Figure 4 3 Projecting yield for timber supply analysis volume table yield curve volume m3 ha 0 00 0 00 300 0 00 0 00 en 7 20 200 41 44 74 60 u 106 24 100 50 336 94 0 336 30 O 50 100 150 200 250 300 350 stand age yr How Is PEM Used with SIBEC to Estimate Site Index Timber supply analysis requires an estimate of site index for every hectare
72. ing and biophysical habitat mapping the history of ecosystem mapping in British Columbia Predictive Ecosystem Mapping User s Guide Page S2 1 What Is PEM 2 What Is PEM 2 1 Background 2 1 1 The Biogeoclimatic Ecosystem Classification System Predictive ecosystem mapping in British Columbia is based on a conceptual model known as the Biogeoclimatic Ecosystem Classification BEC system which was developed in British Columbia and is widely used as a framework for resource management and scientific research The following excerpt of material from the BEC website introduces the reader to the BEC system The BEC system groups ecosystems at three levels of integration regional local and chronological At the regional level vegetation soils and topography are used to infer the regional climate and to identify geographic areas that have relatively uniform climate These geographic areas are termed biogeoclimatic units At the local level segments of the landscape are classified into site units that have relatively uniform vegetation soils and topography Several site units are distributed within each biogeoclimatic unit according to differences in topography soils and vegetation At the chronological level of integration ecosystems are classified and organized according to site specific chronosequences To do this the vegetation units recognized for a particular site unit are arranged according to site history and su
73. irch Horsetail Bedrock mixed Fluvial K Redrock modly calcar Lacustrine cous materials March 28 2005 PEM Critical issues for using PEM in management for biodiversity objectives e PEM accuracy and reliability e Synthesis of the BEC site series classification into realistic planning units Predictive Ecosystem Mapping User s Guide Page S4 15 PEM Applications 4 3 4 Critical Issues for Using PEM to Help Manage for Biodiversity Objectives The critical issues around using PEM as part of an analysis to assist in meeting biodiversity objectives are PEM accuracy and reliability Applying biodiversity objectives at the landscape unit level based on unreliable or inaccurate ecosystem and structural stage mapping could result in the biodiversity plan having erroneously interpreted objectives poor locations for forest ecosystem networks or inappropriate stand structure objectives Synthesis of the BEC site series classification into realistic units for planning The detail in PEM is often generalized by grouping site series into units with similar general ecosystem attributes appropriate for similar biodiversity objectives Scale of PEM mapping relative to the level of objectives PEM is generally produced for 1 20 000 or 1 50 000 scale output March 28 2005 PEM CE Limitations to using PEM with rare or special ecosystem inventories are in the res
74. ld be harvested at any given time Timber supply analysis is usually undertaken with a computer model Regardless of the model used and there are many to choose from timber supply analyses always require the same basic types of input data March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S4 2 PEM Applications landbase forest inventory timber growth and yield and management practices and parameters PEM information may be used in preparing the first three types of data see Section 4 1 4 4 1 3 Modeling Timber Supply Timber supply analyses always follow the same basic procedure regardless of which model is used After the initial gathering of data and documentation of assumptions the analysis proceeds as shown in Table 4 1 Table 4 1 Steps in timber supply analysis to support AAC determination Step 1 The productive forest in a management unit is separated into the Categorize the portion that is available for harvesting the timber harvesting landbase landbase and the portion that is unavailable or inappropriate for timber production Step 2 If management practices are not the same everywhere in the Assign management unit the unit is divided into zones in which different management management objectives requirements and practices can be zones modeled e g watersheds wildlife habitat areas recreation areas Step 3 Forest stands are classified into units that are suitable for projecting Class
75. lysis Branch has published standards for PEM accuracy assessment in TSR PEMs that have very few site series groups and few if any tied equally likely map entities mapped in polygons can be used in a fairly straightforward manner in conjunction with SIBEC estimates in timber supply modeling PEMs with several to many site series groups more than 1 pair per subzone variant or many tied site series mapped greater than 1076 of polygons with tied site series cannot as a rule be used in timber supply modeling It however the assignment of site index estimates to grouped and tied map units in a generalized PEM is reasonable understandable and sensible then the analysis may be acceptable to the Forest Analysis Branch If the PEM is rejected for timber supply analysis purposes e g because it failed the accuracy assessment test there may be alternative ways of estimating site index e g the site index adjustment procedure Spatial Referencing The debate about the significance of spatially explicit timber supply analysis is mentioned here simply because it frequently distracts people from more important issues related to timber supply analysis One of the purposes of tactical and operational planning is to ensure that activities are scheduled to satisfy management requirements for block adjacency Doing so requires a model that can recognize when treated areas are adjacent and then constrain operations accordingly This ab
76. methods of estimating site index based on ecological attributes of the site These methods include the Site Index BEC SIBEC studies and the two approaches veteran trees paired plot used in the Old Growth Site Index OGSI studies The SIBEC studies estimate site index at the site series level of the BEC system The SIBEC tables relate site index to BEC site series for coniferous tree species in British Columbia 26 Clutter J L J C Fortson L V Pienaar G H Brister and R L Bailey 1983 Timber management a quantitative approach John Wiley amp Sons Toronto Ont 77 Carmean W H 1975 Forest site quality evaluation in the United States Adv Agron 27 209 269 28 British Columbia Ministry of Forests 1997 Site index estimates by site series for coniferous tree species in British Columbia Res Br Victoria B C 2001 SIBEC sampling and data standards Version 5 1 Res Br Victoria B C 29 Mah S and G D Nigh 2003 SIBEC site index estimates in support of forest management in British Columbia B C Min For Res Br Victoria B C Tech Rep 004 March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S4 4 PEM Applications Table 4 2 Site index estimates derived from PEM site series and SIBEC are used in the first four steps of timber supply analysis Step 1 Categorizing the landbase Step 2 Assigning management zones Step 3 Classifying the forest inventory Step 4 Pr
77. metric all human made features such as roads buildings and fences as well as natural features such as streams lakes and swamps March 28 2005 se lt L Predictive Ecosystem Mapping User s Guide Page S5 3 Data Types and Attributes Toponymy otficial place names such as Prince George Capilano River etc All models use the DEM and water features from the planimetric coverage PEM standards require that the water features of the PEM be derived from the TRIM coverage Figure 5 2 this ensures that the PEM output is registered to the TRIM water Other data from TRIM can be used during the PEM process for activities such as geo rectification of typed air photos creation of plot files etc Users should understand the strengths and limitations of the TRIM I and TRIM II coverage in their project area TRIM I mapping completed in 1996 produced maps at a scale of 1 20 000 The cartographic framework for this mapping is the Universal Transverse Mercator UTM coordinate system based on NAD83 1983 North American Datum The dimensions of each map sheet are 12 minutes of longitude by 6 minutes of latitude TRIM I data have been updated in many areas of the province TRIM IIS improves upon TRIM I but TRIM II also has some limitations that should be recognized by the PEM model In some areas of the province TRIM II overestimates the amount of water on the landscape for instance in some stream networks produced b
78. nd the ecosystem map Table 4 4 Ecosystem structure based wildlife habitat ratings for old growth dependent species Structural stage Food Security Nesting Shrub herb LOW Nil Nil Pole sapling Moderate LOW LOW Young forest Moderate Moderate Moderate Mature forest High High High Old forest High High High A PEM database can contain up to three site series for every PEM polygon The habitat ratings database houses the capability ratings by animal activity and season of use in the look up table The rows of the look up table represent the BEC variant site series and structural stage and the columns of the table rate the utility of that site series and structural stage to the seasonal use by the animal of interest The ratings go from 1 through 6 with 1 representing the optimal habitat for that species and 6 representing a habitat of no use to the species for that activity and season Joining the PEM site series and structural stage databases to the wildlife capability and suitability ratings tables results in a map of wildlife habitat capability or suitability Figure 4 4 March 28 2005 PEM Habitat models are an abstraction of knowledge about habitat relationships They represent and process this knowledge to predict future conditions Jones et al 2001 Predictive Ecosystem Mapping User s Guide Page S4 10 PEM Applications Figure 4 4 Wildlife habitat capability for ungulate winter range capability is rated
79. nd the second part describes either the relative temperature Interior zones or relative continentality Coastal zones For example the ICHmc stands for the Moist Cold subzone of the Interior Cedar Hemlock Zone Subzone names are abbreviated as letter codes Table 2 1 Table 2 1 BEC subzone names and codes Name Code FIRST PART very dry Relative Precipitation dry moist wet lt Z 3 Q x very wet SECOND PART hot Relative Temperature warm or Continentality mild cool cold very cold hypermaritime maritime 0 3 lt O x 3 lt gt submaritime Coastal Douglas fir CDF Coastal Western Hemlock CWH and Mountain Hemlock MH Zones only Biogeoclimatic variants are given geographic names reflecting their relative location or distribution within the subzone For example the Interior Douglas fir Dry Cool Subzone IDFdk has four variants Thompson Variant Cascade Variant Fraser Variant and Chilcotin Variant Variant names are given number codes e g ICHdk1 which in most cases reflect their geographic distribution within the subzone from south to north Forested site associations are named using one or two tree species followed by one or two understory plant species present in the climax or late seral vegetation unit plant association on which they are based While the species chosen for naming the site association are often abundant in the climax vegetation less common but characteristic March
80. ned both March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A7 3 Appendix 7 systems of naming site series because the MOE two letter codes describe site series for some non forested ecosystems to which MOF codes have not been assigned MOF code is 00 A list of the acceptable two letter codes can be obtained from the Ministry of Sustainable Resource Management at the following web site http www for gov bc ca hre becweb standards becdb htm Are the provincial PEM standards being revised Yes Revised standards are expected in 2006 March 28 2005
81. nfluence the suitability of input data sources for PEM and 2 attributes of PEM outputs for business applications These attributes of data quality have three basic components content precision and accuracy and apply to both thematic and spatial data Table 6 1 outlines the limitations of spatial data sources used in PEM The PEM model relies on input data to generate spatial and database outputs Limitations of the input data must be realistically associated with the consequent limitations of the output PEM data and the applications made with that data Although the spatial precision and accuracy of a PEM may be low it nevertheless may predict the composition of site series in relatively small to large areas This section identifies the limitations of spatial data sources and discusses examples of appropriate and inappropriate uses of PEM data Section 6 discusses the main data sources for PEM Big BEC TRIM Bioterrain Targeted Terrain Forest Cover or VRI and field plot data Other PEM projects use a variety of map sources including geological mapping satellite imagery and soils mapping Table 6 1 Limitations of spatial data used in PEM Data source Spatial limitations Other data Consequences limitations TRIMI DEM generalized lacks Data source is 1 60 000 PEM predictions may miss site series controlled by micro topographic photography strong micro topography and predict an intermediate details Errors in wetland site series rathe
82. nformation in combination with bioterrain or targeted terrain mapping can improve the predictive ability of the PEM model Errors in the TRIM DEM and stream network will be reflected in the output from the landscape facet model Figure 5 6 46 MacMillan R A M V Ketcheson T Robertson K Misural J Shypitka 2003 Canim Lake PEM Final Project Report Unpublished report to Weldwood of Canada March 28 2005 Predictive Ecosystem Mapping User s Guide Page S5 7 Data Types and Attributes Figure 5 6 Landscape facets valley floor 5 slope fine texture crest 45 slope warm aspect coarse texture upper slope 26 slope warm aspect valley floor medium texture 0 BOE organic 5 1 5 Soil Moisture Models Some PEM models use the TRIM DEM and stream network to predict soil moisture regimes using a triangulated irregular network TIN These attributes can help predict the amount of moisture available on the site and the site series that occur there Figure 5 7 Errors in the TRIM DEM will be reflected in the result of the soil moisture model Figure 5 7 Soil moisture predictions AB quet mem 0 55 dry 55 85 NN Ss 7 5 MN 75 255 mu 5 9 5 7 8 5 10 5 EN 105 115 EN 115 125 EN 12 5 20 wet ES No D ata Source LandMapper Environmental Solutions and Weldwood of Canada March 28 2005 PEM h Forest cover and VRI polygon data are often used in PEM models to help pre
83. nt the surface Unlike a grid the TIN allows dense information in complex areas and sparse information in simpler or more homogeneous areas cem ZINN V An example of a triangulated irregular network TIN Variant BEC A subdivision of subzone based on differences in regional climate There can be considerable climatic variability within subzones subzones may be further subdivided into areas that are slightly drier wetter snowier warmer or colder within the subzone Wildlife habitat ratings a classification system describing the ability of a habitat to meet the life requisites of a species or group of species March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page A3 1 Appendix 3 References Appendix 3 References British Columbia Ministry of Forests 1997 Site index estimates by site series for coniferous tree species in British Columbia Res Br Victoria B C 2001 SIBEC sampling and data standards Vers 5 1 B C Min For Res Br Victoria B C B C Ministry of Forests and B C Ministry of Environment 1995 Biodiversity guidebook Forest Practices Code of British Columbia Victoria B C Forest Practices Code guidebook B C Ministry of Forests and B C Ministry of Environment 1995 Riparian management area guidebook Forest Practices Code of British Columbia Victoria B C Forest Practices Code guidebook http www for gov bc ca tasb legsregs fpc fpcguide riparian
84. odels use a raster analysis of the TRIM DEM to identify physiographic features based on landform position relief class slope length slope gradient and inferred hydrologic flow Predictive Ecosystem Mapping User s Guide Page S5 6 Data Types and Attributes 5 1 4 Landscape Facet Modeling Landscape attributes important to the differentiation of site series within a PEM model can be modeled rather than mapped using data from the TRIM DEM Some site series predictably occur in certain landscape positions Figure 5 5 Some PEM models use a raster analysis of the TRIM DEM to identify physiographic subdivisions based on the size and scale of the landform The subdivisions interpret landform position relief class slope length slope gradient and inferred hydrologic flow These attributes can help predict the amount of moisture available on the site and the site series that may occur there Other important landscape settings such as local climatic variation e g frost pockets can also be predicted where these are considered important to the differentiation of site series within a BEC subzone or variant Figure 5 5 Landscape profile most site series occur in predictable positions in the landscape SBSdwl Landscape Profile Soil Particle Size South 4 North Aspect E Aspect we Appendia 10 for defimirians Tree Symbols ser Appendix 2 for definitions ll 1111 M site series number Source Province of B C This i
85. ojecting stand yields One category of land that is not available for harvesting is land of low productivity sometimes referred to as ow site They are considered unsuitable for harvesting and regeneration and therefore excluded from the THLB Site index is often used in the landbase netdown procedure to identify low site areas e g hectares with site index 6 m Since PEM is used to assign site index which in turn is used to identify low site hectares PEM comes into play in this early stage of preparing the data for timber supply analysis If ecosystem type or site productivity is used to identify management zones a PEM inventory can help with zoning the landbase either directly through site series or indirectly through site index If a PEM inventory is used to predict site series and estimate site productivity it will directly influence the classification of the forest inventory Since species and site productivity are the two most important parameters in the growth and yield models used to project yields for each analysis unit the inventory classification must be based on them The most important application of PEM inventories in timber supply analysis is to estimate site productivity or site index for use in projecting stand yields Site productivity is a key parameter in the yield models used to project stand yields Figure 4 3 The yield tables affect the estimate of timber supply both directly and indirectly Since t
86. olution and accuracy of the PEM Predictive Ecosystem Mapping User s Guide Page S4 16 PEM Applications 4 4 Rare and Special Ecosystem Inventories Rare ecosystems are defined as ecosystems that are listed as red or blue by the Conservation Data Centre of the Ministry of Sustainable Resource Management Special ecosystems are those that may be considered rare or unique uncommon interesting sensitive or of concern PEM output and or modeling approaches have been used to predict the spatial location of rare or special ecosystems Figure 4 6 with varying degrees of success depending on the quality of the input data 38 Many of these ecosystems occur within small areas and are difficult to model This activity should not be confused with the Sensitive Ecosystem Inventory SEI which requires air photo interpretation and specialized forms of terrain mapping SEI procedures are generally not applicable to PEM modeling activities Limitations to using PEM with rare or special ecosystem inventories are in the resolution and accuracy of the PEM The input data may not provide enough information to predict the location of these areas The reliability and scale of input data must be carefully considered before using PEM for this interpretation 37 http srmwww gov bc ca cdc 38 Ketcheson M T Dool and and C Littlewood 2002 Development of procedures for mapping special ecosystems in the Arrow Timber Supply Area Unpublished r
87. omprises both the set of data tables mapping the relationships between ecological attributes and site series and the algorithms used to interpret the data and relationships to predict site series The knowledge base can be developed in various ways including using expert knowledge analyzing plot data relative to input variables or combinations of the two Field Data Collection There are at least three reasons for collecting field data to develop the model used to predict ecosystems to test the ability of the model to predict ecosystems from correct input data totest the accuracy of the model s predictions predictive errors may be the result of inaccuracies in the predictive model and or inaccuracies in the input data Field data must be collected to assess model accuracy but is not necessarily required for developing the PEM model which can rely 16 Moon D E and C J Selby 1989 Land resources inventory of the Power River Watershed Land Resource Research Institute Agriculture Canada CEF Ottawa Ont Agriculture Canada Misc Publ No 84 29 17 Moon D E D Dunlop K Iles N Phillips 1999 Problem analysis on data quality assessment issues Draft Submitted to TEM Alternative Task Force by Core Design Technologies Inc March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 5 How PEM Works on existing data expert opinion or data collected in support of the PEM model 5 PEM Cla
88. on a scale of 1 optimal to 6 no value Moose in Winter 18 Elk in Winter Source Province of B C Habitat Supply Models Habitat supply models predict habitat quantity quality and distribution geographically and temporally They vary and there are no standards for their use format and output They are considerably more complicated than wildlife habitat ratings based models although certain habitat supply models may incorporate wildlife habitat capability or suitability ratings into their input data sets A habitat supply model can be used at both the coarse and fine filter levels The coarse filter model can consider landscape level habitat and or ecosystem level habitat and or stand level habitat The fine filter model can consider at its most detailed level the single species guild specific life requisites A habitat supply model may embellish the result of habitat ratings derived from an ecosystem map or it may draw upon the ecosystem map alone to derive a result based on a combination of factors Habitat supply models will often incorporate the results or inputs to a timber supply model 35 Jones K R Ellis R Holt B MacArthur and G Utzig 2001 A strategy for habitat supply modeling for BC Draft Volume 1 Prepared for Habitat Supply Modeling Steering Committee March 28 2005 PEM Critical issues for making PEM based interpretations for habitat analysis accuracy and reli
89. ort Maps Data and Meta data March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 4 How PEM Works The requirements analysis should include a clear statement of inventory objectives a problem analysis to ensure that it is possible to meet the objectives and a clearly defined set of procedures to meet those objectives The knowledge base is the formal expression of the relationships between spatial input data variables and the site series classification Requirements Analysis including Alternatives Assessment Standard inventory procedures cannot answer all questions 1 When initiating a PEM project the client should determine whether PEM is the inventory needed to drive the required interpretations Most important in the requirements analysis are a clear statement of the client s inventory objectives a problem analysis to ensure that it is possible to meet the objectives and a clearly defined set of procedures to meet those objectives Data Assembly and Assessment Spatial inventory data must be assembled and assessed for relevance to the PEM model Evaluation of data quality is an important issue that should be considered early on along with inventory procedures that affect data quality and documentation of the input data 17 Knowledge Base Development In general the knowledge base is the formal expression of the rules used to predict site series from ecological attributes it c
90. ortant for differentiating site series In targeted terrain mapping terrain attributes that are not useful to the model are not mapped Predictive Ecosystem Mapping User s Guide Page S5 5 Data Types and Attributes Bioterrain mapping is expensive and time consuming but could be a valuable stand alone inventory with potential for use in VRI Bioterrain mapping is not a surrogate for Terrain Stability Mapping which gives detailed information for some areas and general information for others Targeted Terrain or Exclusions Mapping Some PEM models use mapping that depicts only the terrain attributes important for differentiating site series Figure 5 4 Terrain attributes that are not useful to the model are not mapped This type of inventory is useful only for the PEM model but is inexpensive and can be completed quickly Targeted terrain mapping increases the ability of the PEM model to predict site series in situations where landscape shape and hydrology as derived by the DEM do not help the model differentiate between site series that occur in similar landscape positions Targeted terrain polygons are generally simple polygons that do not describe attributes through proportions Figure 5 4 Targeted terrain mapping delineates only the terrain attributes which help differentiate site series Source JMJ Holdings Inc and Okanagan Innovative Forest Practices Society March 28 2005 PEM Landscape Facets Some PEM m
91. p www for gov bc ca hre becweb subsite map Bigbec pdf March 28 2005 One of the two critical inputs for a PEM model is a spatial inventory known as Terrain Resource Information Management TRIM Predictive Ecosystem Mapping User s Guide Page S5 2 Data Types and Attributes Figure 5 1 Big BEC in the Cariboo Forest Region Big BEC provides local climatic variation at operational scales EEEE EEEELH es EET 32882 ds Lien Een Quesnel PEM i 3i TE TREE AZ la Br Ae Easter Villas Lake RT 2 y 2005 REM 4 a UT C nim Lake PEM PD002200 HE H sas Pol 1 d Source Province of B C 5 1 2 TRIM Terrain Resource Information Management TRIM is a spatial inventory that covers the province The TRIM program produces digital maps which is a collection of coverages to conform to the BC Geographic System layout A TRIM digital map contains positionally correct complete edited map data for a single map sheet The data consist of all digital planimetric data compiled directly by stereo compilation TRIM contains the following data that may be used in PEM models Digital elevation model DEM representation of the earth s surface as an array of elevations sampled at regularly spaced intervals e Raw contours at a 20 m interval derived from the DEM file e Non positional elevational control points bridge deck and tower heights etc Plani
92. pping is not a prerequisite to PEM Some PEM models use bioterrain mapping as a spatial input data source Bioterrain mapping gives polygon vector format information about terrain materials and their texture origin depth slope relevant geological processes aspect and drainage Bioterrain variables are allocated within a complex polygon Figure 5 3 the location of bioterrain attributes within the polygon is not known Some PEM models use the bioterrain polygon shape as the basis for the final PEM output Most models that use bioterrain use only a few of the variables housed within the bioterrain database The most commonly used variables from bioterrain describe materials that are atypical such as very shallow materials coarse materials fine materials very wet materials organic deposits active fluvial processes or avalanche processes Figure 5 3 Bioterrain map polygons T MISSA ST Yu d TURA E Bioterrai l AN AJO 7 LMC materials n AO d Wes b Z h thickness Qu ha CU Y as j Oe N aspect S CM Lr Ji drainage ys URSS f ora N ER process disturbance NNNM Jc e NN h e R 4 HIN y We Bioterrain is a surrogate for Soil moisture Soil nutrients ET F Ay A EAT Xj RJ A RA PS Source JMJ Holdings Inc and Okanagan Innovative Forest Practices Society March 28 2005 se t lt Some PEM models use mapping that only depicts the terrain attributes that are imp
93. r s Guide Page S4 7 PEM Applications 4 2 Wildlife Habitat Interpretations 4 2 1 Introduction Ecosystem maps generated by PEM can be used to assist in the determination of wildlife habitats Ecosystem mapping provides a consistent framework for looking at habitats of individual species or groups of species at both the coarse and fine filter levels PEM output can be used in rating wildlife habitat capability and suitability of individual species which are based on the relative value of an ecosystem to seasonal life requisites for that species habitat supply models to assist in the analysis and forecasting of habitats and ecosystems to inform decisions about sustainable resource use delineation of ungulate winter range UWR Identified Wildlife Management Strategy accounts and measures for species and their habitats many of which can be determined with the assistance of an ecosystem map generated by PEM 4 2 2 Overview of Wildlife Habitat Assessment Ungulate Winter Range Ungulate winter range UWR is an area that contains habitat needed to meet the winter habitat requirements of an ungulate species Identifying UWR is based on current understanding of ungulate habitat requirements in winter as interpreted by the Ministry of Water Land and Air Protection regional staff based on current literature local knowledge and other regional expertise UWR is recognized under the Forest and Range Practices Act FRPA for t
94. r than the wetter and drier site series shallow lakes and urban produced by the micro topography Unless the micro delineation and topography is strong the model can predict the correct classification site series for the area but the spatial error of the small polygons can be tens of metres 4 Subdominant site series must be included proportionally as an element within the dominant site series based upon expert knowledge of micro topographic elements in the landscape 48 Moon D E 1999 Problem analysis on data quality assessment issues Draft Submitted to TEM Alternative Task Force by CDT Core Design Technologies Inc 49 Moon D 2005 Personal communication March 28 2005 PEM Data source TRIM II Big BEC Bioterrain Targeted Terrain Forest Cover VRI Geological mapping Soils mapping Satellite imagery Field plot or transect data Predictive Ecosystem Mapping User s Guide Limitations of PEM Spatial limitations DEM generalized lacks micro topographic details 10 m contour intervals based on 1 60 000 photography May not accurately reflect BEC variant breaks in all areas when BEC lines are modeled from TRIM contours Complex terrain polygons do not locate individual components Source orthophotos vary in quality and accuracy of orthorectification Spatial location of polygons may be suspect Spatial location of polygons may be suspect Scale is too small to provide a
95. raster and vector based models are largely esoteric From the user s perspective the most important issue is that the end product be usable Conversion between raster and vector formats is simple and comprises an insignificant portion of project costs The most obvious difference is the appearance of map boundaries Figures 3 1 and 3 2 Lines in raster maps are jagged or pixelated i e individual pixels or cells are apparent to the viewer If smooth curves are important spline functions can be used to transform raster maps into smooth curved vector maps with no measurable loss in spatial accuracy 15 Meta data or data about data describe the content quality condition and other characteristics of the data of interest March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 2 How PEM Works Raster and Vector Data Models The rules used to convert real geographical variation into discrete objects is referred to as a data model a set of guidelines for organizing the data in a database The two main types of data model used in GIS are raster and vector In a raster model the mapped area is divided into a regular grid of cells in a specific sequence usually row by row from the top left corner Every location in the mapped area corresponds to a cell in the raster Discrete objects are represented as single cells or groups of related cells In a vector model the location of a discrete point in the mapped ar
96. re began in the 1950s Terrain mapping systems were also being developed standardized and implemented in British Columbia in the 1970s and 80s Nuretin Keser of MOF Research Branch pioneered the conceptual information overlay approach eventually used in EcoGen and ELDAR His approach was based on two seminal publications in the evolution of PEM Hans Jenny s Factors of Soil Formation and Jack Major s concept of adaption to ecosystems soil and vegetation f climate relief parent material organisms time Ecosystem Mapping Biophysical habitat mapping developed in the 1980s delineated ecosystems using air photo interpretation based on permanent 7 Krajina V J 1969 Ecology of forest trees in British Columbia Ecol West N Am 2 1 1 146 Krajina V J 1972 Ecosystem perspectives of forestry H R MacMillan Forestry Lecture Series Univ B C Cent For Continuing Education Vancouver B C pp 1 11 Krajina V J 1977 On the need for an ecosystem approach to forest land management In Ecological classification of forest land in Canada and Northwestern U S A Can Inst For For Ecol Working Group Univ B C Vancouver B C pp 1 11 8 Meidinger D and J Pojar compilers and editors 1991 Ecosystems of British Columbia B C Min For Special Report Series No 6 330 p Pojar J K Klinka and D V Meidinger 1987 Biogeoclimatic ecosystem classification in British Columbia For Ecol Manage 22 119 154
97. redictive Ecosystem Mapping User s Guide Page S4 12 PEM Applications represented in a large polygon For example if the PEM output predicting map unit X7Y2Z1 is superimposed on a different map input A6B4 how are the cross products calculated and which combinations occur in what proportions if the critical habitat is Z If the PEM produces small simple polygons will the habitat model have the processing capability to use the spatial data A difficulty could arise for example if the processing capability of the habitat model was based on an average polygon of 12 ha but the average PEM output polygon was 2 ha The habitat model might have to be subdivided into several smaller areas for processing when the PEM data are added 4 3 Biodiversity Interpretations 4 3 1 Introduction Biological diversity or biodiversity is the diversity of plants animals and other living organisms in all their forms and levels of organization and includes the diversity of genes species and ecosystems as well as the evolutionary and functional processes that link them Developing a biodiversity conservation strategy that is based on various management strategies for individual species is neither feasible nor effective The impact of forest management practices on many species is unknown and practices that benefit some species are often detrimental to others Recommended instead is the development of an ecosystem management approach that provides suitable h
98. roject area then those data can be summarized and compared with the spatial input data variables If the sampling design was appropriate for the methods of analysis statistics can be generated about the relationships between spatial input variables and the site series classification These statistics can then be applied in the knowledge base to determine site series weighting for each spatial variable or combination of spatial variables This approach can be expensive for large project areas with many BEC subzones and variants if there is much field data collection 3 4 2 Testing Knowledge Base The knowledge base should be tested and improved until a satisfactory result is achieved This iterative process can be undertaken using expert review field data or a combination of both Figure 3 4 Expert Review In this process a knowledgeable vegetation ecologist reviews the PEM model output and provides feedback to the PEM practitioner about the model and knowledge base to improve the output of the model Once a satisfactory level of correspondence to the perceptions and experience of the ecologist is achieved the model is finalized and tested for accuracy with independent data Field Data If the PEM project has field data available for use in building the model then the knowledge base can be tested using those data in various ways If the field data were collected with an appropriate sampling regime then they could be summarized and
99. rom the PEM structural stage and site series databases 4 3 2 Overview The more that managed forests resemble forests that were established from natural disturbances the greater the probability that all native species and ecological processes will be maintained PEM output can provide information on both the diversity of ecosystems found and the distribution of structures within those ecosystems at the time of mapping PEM can then be used as a starting point for planning to maintain a variety of patch sizes seral stages and forest stand attributes and structures across a variety of ecosystems and landscapes The juxtaposition of ecosystems and structures found in PEM output can assist in operational planning that maintains connectivity of ecosystems in such a manner as to ensure the continued dispersal and movement of forest and range dwelling organisms across the landscape PEM site series and structural stage output can also be used to assist in the identification of forested areas of sufficient size to maintain forest interior habitat conditions and to prevent the formation of excessive edge habitat Riparian ecosystems can be recognized based on their local extent such that riparian buffers can be maintained and improved based on landscape shape rather than distance buffers 4 3 3 Using PEM in Managing for Biodiversity Objectives PEM output can be generalized into site series groups that summarize suites of related ecosystems that can
100. rs lt lt Combined data matrix Final map PEM output gt 65 Expert knowledge and or data Knowledge base analysis 3 5 Assessing the Suitability of a PEM Inventory Determining whether an existing PEM inventory is suitable for a particular application involves considering the specific requirements of the application and the characteristics of the PEM inventory Tables 3 3 and 3 4 provide an overview of issues that should be considered when deciding whether to use PEM for various applications March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S3 10 How PEM Works Table 3 3 Criteria for evaluating PEM for a specific application Accuracy and reliability Is the PEM sufficiently accurate and reliable for the interpretation Basis of interpretation Is site series appropriate for the interpretation Polygon size and Will the interpretation be constrained by polygon size complexity Polygon complexity Will the interpretation be constrained by polygon complexity Table 3 4 Considerations when specifying PEM attributes for a specific application y e 8 9 i ad 5 om 2 L o Q Y La oa gt Q oY jud G Q E E Hs Q Q zum zum e cu k 0 A o oc 20 53 o 2 I Y icati SE 52 e 38 0 a Application Oz e YS Uc Gv c 9 9 20 20 o ma 9 m 0 z O OE ca o m E Timber Supply Required must Required Both Simple Required Optional Review exceed 65 W
101. s that use a combination of simple rules and fuzzy logic to classify landscape facets spatial representations of combinations of landscape attributes 20 http www for gov bc ca hre ecogen ecoprep htm March 28 2005 PEM Knowledge bases can be derived in different ways depending on the PEM model Predictive Ecosystem Mapping User s Guide Page S3 7 How PEM Works Table 3 2 Part of a knowledge base Thematic input Input layer class Site series number layer description 02 03 04 01 05 06 07 08 Crest 10 Landscape position Lower slope 4 7 17 17 Landscape position Mid slope 1 4 5 Landscape position Upper slope 1 4 2 Landscape position Flat or valley bottom 10 5 7 7 TRIM 0 to lt 5 slope 9 3 10 10 TRIM 5 to lt 25 slope 7 2 2 The way in which a PEM model assigns a site series or combination of site series to a spatial location whether cell or polygon differs with the nature of the model Some models assign a single site series or map unit to a relatively small area others assign site series or map units proportionally to larger areas within an already existing polygon boundary such as a bioterrain polygon or a forest cover polygon Some models determine the site series that wins dominates the cell or polygon and assign that unit to the model s output Others assign a combination of site series to the cell or polygon based upon the first second and third place scores or weights from the knowledge tables
102. sification BEC system Predictive Ecosystem Mapping PEM is a method of ecosystem mapping that relies on computer automated as opposed to manual application of formally structured rules of interpretation the so called knowledge base It usually does not involve air photo interpretation instead polygons are usually derived from other data sources such as forest inventory soils or bioterrain mapping As with TEM the usual PEM mapping unit is the BEC site series If PEM and TEM sound similar it is because the distinction between them is somewhat arbitrary Both TEM and PEM integrate biotic and abiotic ecosystem components of the landscape both are based on expert knowledge both predict occurrence of the ecosystem mapping unit and both allow valuable management interpretations to be made The main difference is that PEM uses a formally structured set of interpretation rules the knowledge base and relies heavily on computer automation whereas TEM uses a less formal set of interpretation rules that are applied manually PEM ostensibly uses existing data rather than air photo interpretation of ecological units and field checking but the A site series encompasses all sites capable of producing the same plant community within a biogeoclimatic subzone or variant at maturity The term predictive is used to indicate that the classification is based on the probability of a particular site series occurring at a specific location
103. ssification The process of using the knowledge base to predict ecosystem occurrence from input data is referred to as classification The spatial inventory data is combined with the knowledge base to produce a spatial depiction of site series and either a separate or linked depiction of stand structure e 6 Internal Accuracy Assessment Quality control is normally Internal and external quality control and assurance procedures and the responsibility of the PEM results should be documented practitioner Quality assurance may be conducted Quality control ensuring that procedures are being followed and Auer neu us quality assurance checking for product quality assess the results as Ma ou being achieved Quality control is normally the responsibility of the an independent third party PEM titi I lit b ducted byth 605555 n practitioner quality assurance may be conducted by the practitioner Internal quality assurance and or an independent third party external quality assurance 7 External Accuracy and Reliability Assessment Interpretations of legislative or policy issues may require assessment of map accuracy and reliability 1819 Note that it is the thematic information rather than the location of polygons linework that is assessed here Where PEM is to be used in the Timber Supply Review TSR an independent assessment of map accuracy and reliability is required see Section 3 6 For use in TSR at least 6576 accuracy o
104. t data report provides general information about the PEM project the area of interest the data used and their sources The reports detail the method of data capture quality assurance quality control for spatial data and any spatial reconciliation undertaken for the input data set as well as any issues of importance relative to the use of that data Table 5 1 presents an example of input data quality assessment Table 5 1 Input data quality assessment example 2 0 TRIM II TNTL Layer Citation TRIM II Program Consultant department Geographic Data Branch Ministry of Sustainable Resource Management Publication scale 1 20 000 March 28 2005 PEM Predictive Ecosystem Mapping User s Guide Page S5 10 Data Types and Attributes Period of compilation Base map projection Quality control Edge matching Edge matching error minimum Edge matching error average Edge matching error max Attribute label matching Raster sized Adjusted control feature shift 1997 Albers NAD 83 As per Table 1and Section 4 4 1 1 in the Standards for Predictive Ecosystem Mapping in British Columbia Inventory Standard RIC 1999 the neatline features in the TRIM II data were considered to be of adequate quality to complete a PEM project at a scale of 1 20 000 No edge matching was required 0 m shift along the x axis O m shift along the y axis 0 m on the xaxis O m along the y axis 0 m on the xaxis O m
105. tive floodplain floodplain plant species and wetlands Archaeological Inventory Guidelines require the description and use of BEC variant vegetation and forest cover and terrain mapping as part of the pre stratification of the landscape for survey planning Interpretations for pre stratification of landscape elements suitable for human habitat using PEM output may be appropriate if designed by knowledgeable personnel Appropriate and inappropriate uses of PEM data for natural resource management Inappropriate use of PEM data Identifying locations of individual site series within complex polygons cutblocks Carnivore Capability Suitability Fur bearers Capability Suitability Species dependent upon small niche habitats Some activity specific habitat in micro topographic positions Stand structure interpretations for meeting biodiversity objectives that are based on a simplistic PEM structure model Deriving stand structure interpretations from erroneous forest cover or VRI data Prediction of rare or special ecosystems that appear in micro topographic locations beyond the resolution of the TRIM DEM Overuse of small scale geology mapping in prediction of rarity based on bedrock geology Deriving rarity or special ness interpretations from erroneous forest cover data Predicting specific vegetation structures coarse woody debris potential stream channel morphology etc are beyond the scope of PEM PEM should not
106. under optimal natural conditions to provide life requisites of a species Complex map entity a map entity that represents more than one site series It is used in situations where the PEM model cannot differentiate between two site series that occur together in a mosaic of micro sites or where individual site series occur in very similar landscape positions Data quality data quality attributes refer to 1 attributes of data that influence the suitability of input data sources for predictive ecosystem mapping and 2 attributes of predictive ecosystem mapping outputs for business applications These attributes of data quality have three basic components content precision and accuracy and apply to both thematic and spatial data Decile the proportion of each site series within a polygon using increments of 1076 and represented by a number between 1 and 10 such that 1 10 of the polygon and 10710076 of the polygon Digital elevation model DEM a representation of the earth s surface as an array of elevations sampled at regularly spaced intervals Ecosystem mapping the stratification of a landscape into map units according to a combination of ecological features primarily climate physiography surficial material bedrock geology soil and vegetation Entity a class or type of thing involved in the mapping process An entity may be a site site series a soil a polygon a plant community et cetera March 28 2005
107. upon the concepts of habitat capability and suitability Capability is the ability of the habitat under optimal natural conditions to provide life requisites of a species Suitability is the ability of the habitat in its current condition to support the species The capability and suitability ratings are developed through literature review field 33 B C Ministry of Water Land and Air Protection 2004 Accounts and measures for managing identified wildlife Introduction V 34 Resources Inventory Committee 1997 Standards for wildlife habitat capability and suitability ratings in British Columbia Terrestrial Ecosystems Task Force Ecosystems Working Group Wildlife Interpretations Subcommittee Victoria B C March 28 2005 PEM Joining the PEM site series and structural stage databases to the wildlife capability and suitability ratings tables results in a map of wildlife habitat capability or suitability Predictive Ecosystem Mapping User s Guide Page S4 9 PEM Applications surveys and local expertise This information is summarized in a species account that identifies the species needs for each season of the year and includes reference to food shelter and reproduction Table 4 4 The information in the species account is related to the BEC classification for the area of interest through a look up table Ecosystem mapping can then be used to spatially depict the habitat ratings by combining the look up table a
108. w of PEM inventory applications Ecosystem mapping is generally conducted to provide a basis for interpretations related to ecosystem based or sustainable natural resource management strategies PEM output is very simple the PEM spatial coverage usually depicts BEC variant site series or map unit which could be a suite of site series and may include modifiers for slope and aspect The stand structure coverage of the PEM model can vary in format depending on the interpretations for which it may be used it usually identifies the dominant and secondary structural stages bryoid grass forb shrub pole sapling young forest mature forest old forest PEM has been used with varying success in several applications in British Columbia these are covered in more detail in Section 4 site productivity assessment for timber supply analysis including Site Index Biogeoclimatic Ecological Classification SIBEC Section 4 1 wildlife habitat interpretations Section 4 2 biodiversity interpretations Section 4 3 rare and special ecosystem inventories Section 4 4 riparian management area interpretations Section 4 5 archaeological overview assessments Section 4 6 March 28 2005 PEM 5 On completing this section users will understand e the differences between raster and vector based PEM modeling and their significance to the final PEM inventory e the framework and concepts underlying PEM
109. y TRIM II a stream is allocated to every gully in areas where streams do not exist Figure 5 2 TRIM elevation contours and water provides Slope class aspect class landscape position hydrologic flow etc MESS The TRIM digital SS X elevation model DEM SSE Source B C TRIM data 44 pritish Columbia Specifications and Guidelines for Geomatics Content Series Volume 3 Digital Baseline Mapping at 1 20 000 Release 2 0 January 1992 45 The new specification for TRIM II 1 20 000 and 1 10 000 Revision Data Capture Version 2 0 May 15 1997 March 28 2005 Terrain mapping can be a useful input to PEM if it is at an appropriate scale and registered to TRIM planimetric features Bioterrain mapping is not a prerequisite to PEM Predictive Ecosystem Mapping User s Guide Page S5 4 Data Types and Attributes 5 1 3 Terrain Mapping Terrain mapping which delineates and classifies surficial materials landforms and geomorphological process can be a useful input to PEM if itis at an appropriate scale and registered to TRIM planimetric features Variations of terrain mapping include bioterrain mapping and targeted terrain mapping sometimes referred to as exclusions mapping Bioterrain Mapping Bioterrain mapping uses the techniques and principles of terrain mapping and also delineates aspect and soil moisture classes in the surficial materials landforms and geomorphological processes Bioterrain ma

PEM Predictive Ecosystem Mapping User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents