GIS Data Management Systems for Territorial Raptors · edit review no_sptl dup_sptl fk12,i13 own_class fk12,i13 owner pvtown_other fk4,i5 state fk4,i5 county usfs_r usfs_f usfs_d

GIS Data Management Systems for Territorial Raptors 1

GIS Data Management Systems for Territorial Raptors Author: David W. LaPlante GIS / Database Administrator Northern California Resource Center [email protected] Abstract Regulatory oversight of Northern Spotted Owls and Northern Goshawks requires data management systems that fully integrate the spatial and tabular data used for analysis and decision support. This paper presents two systems currently used by the U.S. Fish and Wildlife Service in the northern interior region of California and by the U.S. Forest Service throughout Region 5 developed to support those needs. Region 5 Northern Goshawk Database The US Forest Service in Region 5 has been compiling data on the territorial distribution of Northern Goshawk (Accipiter gentilis) for the last several years for the purpose of developing a regional conservation assessment of the species. The principle avian research biologists in the project are Brian Woodbridge, (US Fish and Wildlife Service, Yreka, CA) and John Keane (US Forest Service Pacific Southwest Research Station, Davis, CA). Prior to our involvement in this project, an MS Access database had been developed by John Robinson (US Forest Service Regional Avian Coordinator, Pacific Southwest Region, Vallejo, CA) to provide the necessary data management framework at the time the project had begun. We became involved in this project in the Fall of 2002. At this point, data management needs had evolved significantly since the project had begun, and there was a need to re-develop the data management system. Our first task was to assess those needs and then design and develop a new system to address both immediate requirements as well as a number of additional goals for the project. To begin with, we needed to develop a system flexible enough to:

1) support the continued development of a statewide dataset for the regional conservation assessment and provide for the expanded data needs that had evolved,

2) manage legacy data derived from museum records and historical field notes, as well as currently monitored territories on all land ownerships throughout the State, and

3) support the Northern Goshawk National Monitoring Protocol standards under development by a national team led by Brian Woodbridge, and currently in review at the national level.

In addition, the US Forest Service NRIS Fauna data management system had just been released at version 1.2, and release of version 1.3, which would include “features”,


rather than just survey areas, was scheduled for released in 2003. As NRIS is the corporate standard for the Forest Service, we recognized that the data we were developing would eventually need to be migrated into that system.. With this in mind, we set the additional goals to:

4) develop a system that was as NRIS Fauna compliant as possible, looking ahead at the capabilities to be included in the future, incorporating many of the current data elements and domains as well as a spatially normalized schema that could be reflected in the “features” component of the next release, and

5) use the exercise as a preliminary effort to tease out a data structure and system

logic that, once refined, could potentially be used as a model for the development an NRIS Fauna Goshawk module that could be implemented at a national level.

The primary objective in this project was to develop a system that could adequately model the biology distribution of the species on the landscape, and that was intelligent enough to effectively model the protocol defined methodology for the assessment of an annual status evaluation for those spatially distributed territories. The diagram below illustrates the association of the landscape and survey protocol elements we needed to address in the design.

Conceptual Model of Northern Goshawk Territory Biology Conceptual Model of Northern Goshawk Territory Biology and Database Elementsand Database Elements

SURVEYSSURVEYS

PROTOCOLPROTOCOL

INCORPORATESSPECIES

KNOWLEDGE

TERRITORIESTERRITORIES

Spacing rules differentiate many territories within an area of interest

INDIVIDUAL TERRITORYINDIVIDUAL TERRITORY

A territory is a spatial feature defining an area of occupancy on the landscape

Can contain multiple nest site locations

NEST SITESNEST SITES

One or many within an individual territory

Annual status determined by aggregate of nest sites

ANNUAL STATUSANNUAL STATUS

Based on annual surveys

Keyed to individual nest site breeding and occupancy data

Level 3

Level 1Level 2

Applied toIndividualTerritories

Determinationof

Activity


One of the initial tasks was to define the spatial normalization of the data. Northern Goshawk territories are comprised of one or more nest site locations within an area occupied and defended by a breeding pair. This normalization is initially reflected in the tabular data model through a one to many relationship between territory and nest site records. Nest site features are first derived from field surveyed UTM coordinates or legal descriptions. Territory features are then generated from the composite buffers associated with those locations.

Spatial Model on Northern Goshawk TerritoriesSpatial Model on Northern Goshawk Territories

Nest site within a specified distance of one another are

associated with a single territory

Research into the distribution of multiple nest sites within discrete territories and the distribution of territories on the landscape has yielded the following results:

• Mean distance between territory centers = 2.36 miles • Mean distance between alternate nests = 320 yards (max = 2426 yards) • Radius containing all nests within territory = 1500 yards

The associations between nest site locations and territories were defined by the project researchers in their review and compilation of data over the life of the project. Data development proceeded with the research biologist’s review of data from contributing forest and private parties, evaluation of those associations, then entry of data into the system that is subsequently used to develop the spatial features. All data that


populated the legacy database had to be translated into the re-designed schema prior to the delivery of the system to the researchers conducting the project. Perhaps the greatest challenge in the exercise was the need to tease out and translate the biological knowledge of the research biologists into a system that could (1) effectively model the behavior of the species of interest on the landscape and (2) be used to derive an legally defensible Annual Status evaluation for a given territory that could then be used as an adaptive management decision support tool. To that end, we needed to develop a protocol decision tree to outline the rules defined within the draft Northern Goshawk National Monitoring Protocol as it applies to the evaluation of that Annual Status designation. There are significant and subtle nuances in the evaluation process as defined by the biologists that needed to be engineered into the system programmatically. For example, a territory can only be defined as UNOCCUPIED if a comprehensive Level 3 survey encompassing the entire territory has been conducted, where as a Level 1 or Level 2 survey in the proximity of a known nest site can only lead to an evaluation of INACTIVE if the nest is found to be unoccupied. These subtle distinctions are carefully structured into the decision tree, which is navigated top to bottom to derive a final determination Annual Status.

Monitoring Protocol Decision TreeMonitoring Protocol Decision Tree

The system is designed to capture the resource knowledge of the research community at a refined level.

• The draft National Monitoring Protocol defines levels of activity within a territory with a combination of survey parameters and activity values.

• The decision tree is translated into a validation matrix used to control value choices available in the data entry screens.

Determination of nest site and territory status is based the on combination of survey level, survey type, survey application, and survey effort.Determination of activity is based on the combination of the preceding criteria and Occupancy Status, Breeding Status, and Productivity.

Informal Unknown R5 Protocol Informal Unknown R5 protocol

N/A No Data Territory Monitoring

N/A N/A Level 3

N/A

N/A

No Data

N/A

Territory Monitoring

Level 1 Level 2

UN OC

UN

UN

AC

0 – 6Y

OC UN NO

NB UN NA

NA UN NA

NO UNOC

I AC UN

UN

UN

NA 0-6Y UN

OccupancyStatus

BreedingStatus

Productivity

Survey Level

Survey Type

SurveyApplication

Survey Effort

Territory Level Individual Nest Sites

Informal Unknown R5 Protocol Informal Unknown R5 protocol

N/A No Data Territory Monitoring

N/A N/A Level 3

N/A

N/A

No Data

N/A

Territory Monitoring

Level 1 Level 2

UN OC

UN

UN

AC

0 – 6Y

OC UN NO

NB UN NA

NA UN NA

NO UNOC

I AC UN

UN

UN

NA 0-6Y UN

OccupancyStatus

BreedingStatus

Productivity

Survey Level

Survey Type

SurveyApplication

Survey Effort

Territory Level Individual Nest Sites


The development of the decision tree allowed us to define a hierarchal set of data elements that are used to sequentially traverse the decision tree, and forms the foundation of an evaluation engine engineered into the system that programmatically constrains the development of data into a suite of protocol compliant values that resolve into a protocol compliant Annual Status evaluation. The conceptual design of the system is shown in the graphic below. The design needed to (1) reflect the spatial normalization of nest sites and territories, (2) provide for an expanded suite of Fauna compatible data, and (3) incorporate all of the protocol compliant data elements and domains. The heart of the evaluation engine is a validation matrix that manages all valid protocol compliant combinations of values used to derive the Annual Status for a territory. The system uses this matrix to programmatically navigate the decision tree defined above.

CalGos Conceptual Database DesignCalGos Conceptual Database Design

ACTIVITY recordsare accociated with

TERRITORY orNESTSITE TREE

records

EVALUATIONENGINE

TERRITORIES

NESTSITETREES

ACTIVITY

DETECTIONS

INDIVIDUALS

PROTOCOL

Spatially normalized data structureSpatially normalized data structureSurvey Records and associatedSurvey Records and associatedFauna compliant data elementsFauna compliant data elements

Development of protocol specificDevelopment of protocol specificevaluation of annual statusevaluation of annual status

based on decision treebased on decision tree


The logical design of the system is comprised of the primary object classes (territories, nest site locations, observation records) and their associated type tables used to manage valid domains, and the validation matrix and its associated type tables that support the evaluation engine, as well as the suite of supplemental Fauna related data.

CalGos Database Logical DesignCalGos Database Logical Design

tblTERRITORY

PK TERRITORY_ID

U2 TIDNEW_RECORDS

U1 LEG_IDFK1,I1 SOURCE

TERRITORY_NAMETERRITORY_NOYR_DSCVRCDFG_IDNDDB_IDOTHER_IDTERRITORY_REMARKSSTEWARDCREATED_DATE

vldSource

PK SOURCE

I1 IDSOURCE_DESCRCREATED_DATECREATED_BY

tblACTIVITY

PK,FK1,I2 TERRITORY_IDPK NESTSITE_IDPK SRV_DATE

U1 AIDI1 LID_FKI3 TID_FK

SVR_DATE_VALIDSRV_YRACTIVITY_IDCHECKED

FK10,I12 SRV_TYPEFK2,I4 APPLICATION_IDFK8,I10 SRV_EFFORTFK7,I9 OCCUPANCYFK3,I5 BREEDING_STATUSFK6,I8 NUM_YOUNGFK5,I7 GROUP_TYPEFK4,I6 DATA_SOURCE_TYPE

PRIMARY_SURVEYORFK9,I11 SURVEYOR_QUALS_TYPE

SRV_REMARKSCREATED_DATECREATED_BY

tblLOCATION

PK,FK2,I3 TERRITORY_IDPK NESTSITE_ID

U1 LIDI1 TID_FK

TAGGEDEDITREVIEWNO_SPTLDUP_SPTL

FK12,I13 OWN_CLASSFK12,I13 OWNER

PVTOWN_OTHERFK4,I5 STATEFK4,I5 COUNTY

USFS_RUSFS_FUSFS_DBLM_DBLM_RACA_DESNPS_DES

FK10,I11 TWNFK8,I9 RNGFK9,I10 SECFK6,I7 QTRFK7,I8 SIXTEENTHFK5,I6 MERIDIAN

TWN_LEGACYRNG_LEGACYSEC_LEGACYQTR_LEGACYSIXTEENTH_LEGACYMERIDIAN_LEGACYQUAD_NAMEQUAD_NAME_LEGACY_RECALCQUAD_NAME_LEGACYQUAD_NOUTMEUTMNUTMZONEX_COORDY_COORDLONGLAT

FK11,I12 LOC_METHODLOC_ACCURLOC_ACCUR_UOM

FK3,I4 F_COMMUNITYF_STAND_TYPE

FK13,I14 HOW_TYPEDFK16,I17 F_WHR_TFK15,I16 F_WHR_SFK14,I15 F_WHR_D

G_WHR_TG_WHR_SG_WHR_D

FK1,I2 G_KUCHLERHOW_TYPED_LEGACYREVIEW_COMMENTSCREATED_DATECREATED_BY

lutKuchler

PK CLASS_ID

KUCH_IDKUCH_CLAS

SI1 KUCH_CODE

HABITAT

vldCommunityType

PK COMMUNITY_ID

I1 IDCOMMUNITY_DES

CCREATED_DATECREATED_BY

vldCounty

PK,FK1,I3,I2 STATEPK COUNTY

I1 IDCO_ABBFIPSST_FIPSCO_FIP

SLABEL

vldLEGAL_MER

PK MER

I1 ID

vldLEGAL_QTR

PK QTR

I1 ID

vldLEGAL_RNG

PK RANGE

I1 ID

vldLEGAL_SEC

PK SEC

I1 ID

vldLEGAL_TWN

PK TOWNSHIP

I1 ID

vldLocationMethod

PK METHOD_CODE

I1 IDI2 LOC_METHOD

LOC_METHOD_DESCR

CREATED_DATECREATED_BY

vldOwner

PK,FK1,I1 OWN_CLASS_IDPK OWNER

IDSTATEOWN_IDAGENCYOWN_UCASEPOWNCREATED_DAT

ECREATED_BY

vldStandTypeMethod

PK TYPE_METHOD

I1 IDTYPE_METHOD_DESC

RCREATED_DATECREATED_BY

vldWHRDensity

PK WHR_DENS

I1 IDDENS_RANGEWHR_DENS_DES

C

vldWHRSize

PK WHR_SIZE

I1 IDWHRSZ_CL

SWHR_DBH

vldWHRType

PK HAB_TYPE

I1 IDHAB_NAM

E

gisOVERLAYS

PK DB_ID

LIDGIS_ID

FK1,U2,U1 TERRITORY_IDFK1,U2,U1 NESTSITE_ID

OWNERCODEFIPSCOUNTYSTATESYMBOLDESCRIPTIOSCIENTIFICSHADESYMRECLASSHABITATGROUPCLASSIFYMODIFYDESC_IDCLASS_IDKU_RECNOSPECIESWHRCVNAMEWHRNAMETOWNSHIPRANGESECMEROCODEMAPNAMEUSGS100TID_BAK

vldStates

PK STATE_ABBR

I1 IDSTATE_NAMESELECT

vldOwnClass

PK,U1 OWN_CLASS_ID

I1 IDU1 STATE

OWN_CLASS_NAMEGOWNFILTERCREATED_DATECREATED_BY

vldApplications

PK,I1 APPLICATION_ID

I2 IDI3 APPLICATION

APPLICATION_DESCCREATED_DATECREATED_BYSELECT

vldBreedingStatus

PK BREED_CODE

I1 IDBREED_VALUEBREED_DESCOCC_CONSTRAINTSELECT

vldDataSourceType

PK DET_DATA_SOURCE

DECSRIPTIONDET_DATA_SOURCE_C

N

vldGroupType

PK,I1 GROUP_TYPE

SELECTDESCRIPTIONGROUP_TYPE_CN

vldNumYoung

PK,I2 NUM_YOUNG_ID

I1 IDI3 NUM_YOUNG

SELECT

vldOccupancy

PK OCC_CODE

I1 IDOCC_VALUEOCC_DESCSELECT

vldSurveyEffort

PK SURVEY_EFFORT_ID

I1 IDI2 SURVEY_EFFORT

SURVEY_EFFORT_DESCROCC_CONSTRAINTSELECT

vldSurveyorQualsType

PK,I1 OBSERVER_QUALS

DESCRIPTIONOBSERVER_QUALS_C

N

vldSurveyType

PK,I2 SURVEY_TYPE_CODE

I1 IDI3 SURVEY_TYPE

SURVEY_TYPE_DESCRCREATED_DATECREATED_BYSELECT

tblDETECTION_METHOD

PK,FK1,I2 TERRITORY_IDPK,FK1,I2 NESTSITE_IDPK,FK1,I2 SRV_DATEPK,FK3,I4 DETECT_METHOD

U1 DIDI1 AID_FK

CREATED_DATE

tblINDIVIDUALS

PK,FK1,I2 TERRITORY_IDPK,FK1,I2 NESTSITE_IDPK,FK1,I2 SRV_DATE

U1 IIDI1 AID_FK

NG_COUNTCREATED_DAT

E

tblPROTOCOL_METHOD

PK,FK1,I1 TERRITORY_IDPK,FK1,I1 NESTSITE_IDPK,FK1,I1 SRV_DATEPK,FK2,I2 PROTOCOL_METHOD_ID

AID_FKPIDCREATED_DATE

ACTIVITY_CODE

PK,FK7 GROUP_TYPEPK ACTIVITY_ID

I1 ACTIVITY_CODEACTIVITY_DESCNESTSITECHECKED

FK6,I12 SURVEY_TYPEFK1,I6 APPLICATION

PROTOCOL_METHODFK5,I11,I5 N_SURV_EFFORTFK4,I4,I10 N_OCCUPANCYFK2,I2,I7 N_BREEDING_STATUSFK3,I3,I9 N_NUM_YOUNG

COMMENTS

lutParamSrvMatrix

PK,FK6 NUM_YOUNG_IDPK ID

CHECKEDNESTSITE

FK5,I6 SURVEY_TYPEFK1,I1 APPLICATION

PROTOCOL_METHODFK4,I5 SURVEY_EFFORTFK3,I4 OCCUPANCYFK2,I2 BREEDING_STATUS

GROUP_TYPE

vldIndvActivityOther

PK,I1 INDIV_ACTIVITY_OTHER

RANKLOOKUPSELECTDESCRIPTIONINDIV_ACTIVITY_OTHER_

CNCREATED_DATECREATED_BY

vldIndvActivity

PK,I1 INDIV_ACTIVITY

RANKLOOKUPSELECTDESCRIPTIONINDIV_ACTIVITY_C

N

vldIndvAgeType

PK,I1 INDIV_AGE

SEQUENCESELECTDESCRIPTIONINDIV_AGE_C

N

vldDetectMethodOther

PK,I2 DETECTION_METHOD_OTHER

RANKI1 LU

SELECTDESCRIPTIONDETECTION_METHOD_OTHER_

CNCREATED_DATECREATED_BY

vldDetectMethod

PK,I2 DETECTION_METHOD

RANK_BWRANK_LU

I1 LUSELECTDESCRIPTIONDETECTION_METHOD_

CN

Wednesday, February 18, 2004

Page 1

R5 Northern Goshawk Territory Database

Application Development:

Dave LaPlanteNorthern Califrornia [email protected](530)468-2888

Researchers:

Brian WoodbridgeResearch Wildlife BiologistUSDA Forest ServiceKlamath National [email protected]

John KeaneResearch Wildlife BiologistUSDA Forest ServicePacific Southwest Research [email protected]

This group of tables and relationshipscomprise the validation matricesbuilt into the data structure.

These are used as the source recordsetto define valid values returned to the combo box controlsin the application's data entry screens.

This structure also provides the framework for developingan extended logic engine within the systemthat can be used to apply the resource intelligencederived from the research communityto the analysis of data managed within the system.

lutParamGroupType

U1 CHECKEDFK4,U1,I4 OCCUPANCYFK1,I1,U1 BREEDING_STATUSFK3,U1,I3 NUM_YOUNGFK2,I2,U1 GROUP_TYPE

CountOfID

ACTIVITY_CODE is the table used tocrosswalk legacy ACTIVITY_ID values

to the new set of elemental surveyparameter and activity values

Records from tblACTIVITY can be joined to tblLOCATIONSby forcing the system to create and maintain

referential integrity programatically

Spatially normalized object classes and type tables

Fauna compliant data elements and domains

Validation matrix and associated type tables


The evaluation engine is implemented through the interaction of the validation matrix and its associated type tables describe above, and a set of data entry screens that programmatically control and constrain the automation of data into the system.

Data Entry ScreensData Entry Screens1) The switchboard in the entry point to the data entry screens,

pre-built queries, reports, and other application components.

2) The Territory data entry screen is used to automate data for territories and nest site trees.

One or many nest site trees are associated with each territory. All location and ownership information is managed in the nest site tree record.

3) The Survey/Activity sub form displays read only survey data associated with either the territory or the nest site tree. Both discovery survey and territory monitoring information are available here.

4) Existing records can be viewed in detail and accessed for edit using the Vu / Edit button on the sub form.

5) New survey records are automated using the Survey / Activity data entry screen.

6) The Survey / Activity data entry screens are comprised of groups of functionally related fields and sub forms that work in a coordination.

The primary data entry screen manages the entry of information specific to the territory and its one or many associated nest site records, which is the level at which location data is generated. Existing information in the system for surveys associated with the territory or nest sites is displayed in a read only format on the bottom of the primary screen. This section contains controls that allow either (1) entry of new survey records or (2) edit access to existing records. Built in type tables provide validation control of all combo box controls, which are themselves designed to be interactive with other controls on the form, such that value changes in one control will automatically clear out, reset, or constrain values in other controls where dependencies exist. For example, the data for region 5 includes territories in Nevada and Southern Oregon, as well as throughout California. Selecting a state automatically constrains the list of values for Counties to the state selected. If a state and county had been previously selected, changing the value for the state will reset all state dependant values in other controls (county, ownership). Every effort was made to develop a system that precluded the potential for data entry errors, ensuring throughout the data entry process that only research quality data could be developed.


Once territory and nest site data are developed, the user can then enter survey information. The data entry screens used for this are carefully designed to walk the user through a data entry sequence designed to navigate the protocol defined decision tree. Data entry controls are organized into functional groups of dependant elements. There are a group of 9 controls used to navigate the decision tree. Data can only be entered into the system following a discrete sequential order. Successive combo box control queries the validation matrix using the values entered the preceding controls, so that only valid protocol compliant values are presented for selection to the user. If changes are later made to values in controls earlier in the sequence, all subsequent controls in that functional group are cleared out, and the user must re-navigate the decision tree through the data entry sequence. The system tests the controls in sequence for null values and returns notification, forcing complete automation of data, and always moves the cursor into the appropriate control.

Functional Groupings and Dependant ElementsFunctional Groupings and Dependant ElementsTwo levels of functional groupings

1) Sequentially Dependant Combo Box Controls

2) Dependant Sub Form Controls

• Queries the validation matrix against the values in ALL of the preceding controls in the group

• Dispersed within the GENERAL INFORMATION, SURVEY PARAMETERS, and ACTIVITY VALUES sections on the screen.

• Sequentially numbered to guide the sequence of automation.

• Available values for any given control will be different based on the preceding combinations

• The system tests the controls in sequence for null values and returns notification, forcing complete automation.

• These sub forms manage and display related records in each category associated with the current survey record

• #1) TERRITORY OR NESTSITE SURVEYED? must be checked to automate child records.

• If child records exist and the SURVEYED? control is unchecked, the system returns notification that child records must first be deleted.

The dependant sub-form controls are used to automate additional NRIS Fauna compliant information. The ability to enter data into these controls is dependant on data already entered into the preceding set of controls. For example, if the initial control that defines whether or not a survey was conducted at a site (negative data) is not checked,


no data may be entered for individuals observed. If data already exists for individual observations, it must be manually deleted out of the system before the initial control can be unchecked so as to avoid accidental deletion of child records. Again, the system returns notification and guides the user through the data entry process. Throughout the design process, careful consideration was given to the programmatic framework needed to control the development of valid data within the system through user friendly and intuitive interface. In addition to the standard data entry screens, a validation list management interface was developed to provide access to and management of the various type table values that needed to be updateable by the researchers using the system. All peach colored coded controls on the data entry screens have validation lists that can be edited with administrative privileges. The Validation List Management Interface is accessed through buttons in the central control section of each data entry screen. A series of list specific forms can then be accessed through this interface, some of which manage simple lists, others complex lists with multiple dependant data elements.

Validation List Management InterfaceValidation List Management Interface

• The Validations button on either the Territory or Survey / Activity data entry screen opens the interface.

• When opened, the data entry screens are set to not visible to avoid interacting with too many screens.

• Returning to the data entry screens brings them back with the current record available for edit.

1) Access

2) Interface

• Command buttons open separate edit screens for each validation list.

3) Simple data entry screen4) Extended data entry screen

• Two types of data entry screens provided.

• Used to manage simple lists

• Add and Delete values with command buttons

• Select value in list box to populate edit block• Used to manage complex lists

• Select category in Filter combo box

• Distinct selections enable different edit control subsets


If at any time in the data entry process the need to update a list presents itself, the user can move to the Validation List Management Interface, select and edit the required list, and immediately return to the current record in the buffer and have immediate access to the updated list. Again, the system is designed both for ease of use and complete control of data quality. To round out the system, we developed a comprehensive User’s Guide that addresses all details of the working with the system.

User’s GuideUser’s Guide

The Table of Contents is hyperlinkedto bookmarks throughout the document .

The User’s Guide is organized to provide detailed information about the function and use of each component in the system available to the user.

• Navigation is intuitive and efficient

• Provides rapid access to required information.


Additionally, the system was designed to be deployed as a distributed solution, with two independent researchers developing and managing their own data that would be compiled at a later date. The system is programmed to be deployed on a discrete desktop or in a file server environment, and automatically hooks itself up on installation. This system has been in use by the research biologists conducting this study for over a year at this point, and the system has worked flawlessly for them. At this point we have compiled the individual datasets and are currently working with the data in support of the conservation assessment in progress. We are also working with John Robinson and a US Forest Service enterprise team in Region 5 to migrate these data into NRIS Fauna v 1.3. Fauna is built on an elegant architecture that provides the ability to model data flexibly within that system. As a result, we have the opportunity to determine how best to model these data within that system, develop conventions that can assure the consistent development of assessable data into the future, and translate these data into that model for distribution to individual forests throughout the region.


Northern Interior Spotted Owl Geodatabase System The US Fish and Wildlife office in Yreka, CA, is responsible for THP plan review and HCP negotiations with private industrial timberland owners as well as for consultation with the US Forest Service on timber sales within the Northern Interior region of California as they affect endangered species in the region. One of the primary species of regulatory concern to the agency is the Northern Spotted Owl (Strix occidentalis). We have contracted with that agency to develop a geodatabase system to support their needs during the assessment and review process. All timber harvest plans in California are required to be reviewed relative to their potential impact on the Spotted Owl using the state’s official Spotted Owl database, developed and managed by Gordon Gould (BIOS Project, Wildlife and Habitat Data Analysis Branch, California Department of Fish and Game). Gordon has maintained this database for many years, and actively updates the data in his system throughout the year as new data is submitted by timber management entities. Gordon’s database is re-distributed on a regular basis. As a result, one of the primary design goals in this system was to efficiently re-process and output tabular and spatial data reports and map documents when a re-distribution occurs to ensure that Fish and Wildlife biologists are always working with current products. One need of Fish and Wildlife biologists is to have the data in Gordon’s database -processed so that an annual Reproductive Status report for any given territory can be produced. In that multiple surveys may have occurred at any given territory in a particular year, we need to calculate an annual reproductive status based on the observations with the greatest reproductive rating. For example, if multiple survey records exist, the single observation with the highest reproductive status needs to be identified, from which a specific annual status value is calculated and reported. In addition, there is a need to know when surveys were conducted and no owls were observed (negative data) as well as years when no surveys were conducted. This later information is used to identify data gaps in the state database, and assist the service and timberland managers in identifying the need to submit data to the state to fill those gaps. This is especially important in the process of HCP negotiation. Of equal import is information about ownership associated with the 1.3 mile activity center buffers (legally defined territories). This information helps identify the entities responsible for management and survey of these territories, and again is used to identify and fill data gaps to the greatest degree possible. The system we’ve designed to meet these business requirements functions as an integrated MS Access relational database / ArcGIS geodatabase system, and is comprised of multiple independent data processing subsystems that leverage both the SQL capabilities in MS Access and the new ArcGIS 9.0 geoprocessing capabilities. The release of ArcGIS 9.0 has allowed us to fully automate the spatial data processing


and map production, creating significant efficiencies in the management of the workflow associated with this project. Data is initially provided by Gordon as two MS Access tables, the first with one record per territory (or activity center), the other with potentially many survey / observation records for each territory. Historically, observation records extend back to the early 1900s. Processing the data first requires an initial validation review. We accomplish this using a suite of check queries designed to identify data integrity issues. Once validation is completed, the data must be processed into a set of tables with (1) validated referential integrity as well as

(2) valid data types used by the geodatabase system. This is accomplished using a macro in MS Access that runs a sequence of SQL procedures used to recalculate field data types, assure referential integrity, and prepare the data for migration into the geodatabase system. At this point, all activity center records are spatially identified with a set of UTM coordinates.


Once the initial pre-processing is completed, we run an ArcGIS 9.0 geoprocessing workflow model that loads the geodatabase system with all object and feature classes calculated from the state database. The model performs the following functions: 1) First delete then import the updated observation record object class 2) Create activity center features from UTM coordinate XY events through an overlay

procedure with a regional quadrangle dataset to add relevant data to the attribute table

3) Create buffers at 0.25, 0.7, and 1.3 mile radii for various overlays. These buffer

distances define the various management areas associated with each activity center; 1.3 mile buffers define the area of legal responsibility, 0.7 mile buffers define the area of foraging relative to habitat suitability and impacts, and 0.25 mile buffers define the nest core area.

4) Create all relationship classes between the activity center feature class, the

observation record object class, and derived buffer feature classes

Make XYEvent Layer

tb lNSOTERR_Layer

Buffer ACBuf_130

Buffer (2) ACBuf_070

Buffer (3) ACBuf_025

CreateRelat ionship

ClassrelACHasBuf130

CreateRelat ionship

Class (2)relACHasBuf070

CreateRelat ionship

Class (3)relACHasBuf025

CreateRelat ionship

Class (4)re lACHasNOSIGHT4

Add Field ACBuf_130(BufDist)

Add Field (2) ACBuf_070(BufDist)

Add Field (3)ACBuf_025

(BufDist)

CalculateField

ACBuf_130(Calc BufDist)

CalculateField (2)

ACBuf_070 (CalcBufDist)

CalculateField (3)

ACBuf_025(Calc BufDist)

Intersect AC

Make FeatureL a y e r quads_Layerquads.shp

tblNSOTERR

tblNOSIGHT4 (3)Delete Delete

succeededtb lNOSIGHT4

USFWS_NSO.mdb

Table toGeodatabase

(multiple)

USFWS_NSO.mdb(2)

tblNOSIGHT4 (2)


The next step is to then run a second geoprocessing workflow model that runs the entire overlay procedure used to derive spatial statistics required for the MS Access reports that the system currently outputs. This workflow model automates the following procedures: 1) Overlay both activity centers and activity center 1.3 mile buffers on three sets of

parcel layers in the region using discrete spatial query criteria, 2) Calculate new fields and their associated values into each derived output layer that

are essential to the subsequent processing of the data 3) Append the overlay derived datasets into two final datasets; one for activity center

point locations, the other for ownership polygons associated with each activity center buffer,

4) Calculate frequency tables used in the subsequent MS Access SQL sub-system to

calculate the output result tables and reports.

ACBuf_130Make Feature

L a y e rACBuf_130

_ L a y e r

Select LayerBy Location

ACBuf_130_Layer (SA1)


(2)

ACBuf_130_ L a y e r(SA2)

S isk i youParcels

SPI Own

Intersect tmpSA1_ACB_O W N

Intersect (2)tmpSA2_ACB_

OWN

ACMake Feature

Layer (2)AC_Layer

Make FeatureLayer (3) AC_Layer1

Make FeatureLayer (4)

ACBuf_130_Layer1

(SA2)


ACBuf_130_Layer1 (SA1)


spi_own_L a y e r


s is_parcels_Layer

Intersect (3)tmpSA1_AC

_OWN

Intersect (4)tmpSA2_AC_

OWN

Add FieldFC

NewFieldCalculate

FieldtmpSA1_ACB

_OWN (2)

Add Fie ld (2)FC

NewFieldCalculateField (2)

tmpSA1_AC_OWN (2)

Add Fie ld (3)F C


tmpSA2_AC_OWN (2)

Add Fie ld (4)FC


tmpSA2_ACB_OWN (2)


tmpSA1_ACB_OWN_Layer


tmpSA1_AC_OWN_Laye




tmpSA2_AC_OWN_Layer

Append SA_ACB_OWN(Append)

Append (2)SA_AC_OWN

( A p p e n d )

C o p yFeatures SA_ACB_OWN

CopyFeatures (2)

SA_AC_OWN

SPI OwnBoundary


(3)

AC_Layer(Select)


AC_Layer2


(4)

AC_Layer1(Select)


AC_Layer3

JacksonParce ls

Make FeatureLayer (14) taxlots_Layer


(5)

ACBuf_130_Layer

(SA1.1)


ACBuf_130_Layer1

Intersect (5)tmpSA11_ACB_

OWNAdd Fie ld (5)

FCNewField

5

CalculateField (5)

tmpSA11_ACB_OWN (2)



Make FeatureLayer (17) AC_Layer4


(6)

AC_Layer4(Select)


AC_Layer5

Intersect (6)tmpSA11_AC_O

W NAdd Fie ld (6)

F CNewField

CalculateField (6)

tmpSA11_AC_OWN (2)


tmpSA11_AC_OWN_Layer

Frequency SA_ACB_OWNFrequency

Se lec ted F ie lds

Have Their Centers InNew Select ion

Completely WithinNew Select ion

Al l F ie lds

Al l F ie lds

Selected Fie lds

Se lec ted F ie lds

IntersectNew Select ion

Selected Fie lds

IntersectNew Select ion Selected Fie lds

Selected F ie lds

IntersectNew Select ion

Al l F ie lds

IntersectNew Se lec t ion Selected Fie lds


Once this procedure is completed, the final step in re-processing data is run using an MS Access SQL sub-system. This sub-system is comprised of 22 sequential SQL procedures (and associated form and report objects) that ultimately calculate two results tables used as the record source for the final reports generated within the system. Data derived in the overlay procedures are processed through this system to derive the various ownership statistics and spatial report elements. Observation records in Gordon’s data are separately processed and translated into the set of reproductive status values used by Fish and Wildlife biologists to adequately assess biological conditions at the territory level.

uvq_SA_ACB_OWN uvq_SA_AC_OWN

uvq_MasterowlByOwn

q_Obs8702_calc

SA_ACB_OWN SA_AC_OWN tblNSOTERR tblNOSIGHT4

GIS overlay results Gordon Gould's Data

q_Obs8702_All

q_Obs8702

q_Obs_Xtab_ObsCount

q_Obs8702_MaxRating

q_Obs8702_Xtab_RA

q_Obs8702_Xtab_Rpt

q_Obs8702_Group

q_Obs8702_Group_ObsCount

uvq_Obs8702_ObsYrq_Obs8702_Group_OwlYrComb

q_Obs8702_Group_RepStatus

q_Obs8702_Group_RepStatus_mtq

tmpObs8702_Group_RepStatus

q_ACB_PcntOwn_Obs8702_mtq

tmpACB_PcntOwn_Obs8702

ACBuff_13

q_ACB_PcntOwn_calc

q_ACB_PcntOwn

AC

GIS NSO Feature Classes

q_ACB_PcntOwn_Xtab

q_ACB_PcntOwn_Xtab_Rpt

rpt_ACB_PcntOwn_RepStatus

Friday, June 25, 2004

Page 1

USWFS NSO Ownership Analysis SQL Processing Diagram

Access report toaccompany map series

Excel spreadsheetof results

eq_ACB_PcntOwn_Obs

Sub-system used to calculatetmpTable recordset that serves

as the source for the lsit boxcontrol in the final report

defining the unique activitystatus values by year by Master

Owl

q_ACB_PcntOwnClass_Xtab

rq_ACB_PcntOwn_RepStatus_Final frmRptMgr

Access form to definestudyarea for report

mcrObs8702_PcntOwn


Output of the final reports is controlled through a Report Manager interface that allows reports to be generated for a select study area, or for “All” study areas, and for a select ownership or for “All” ownerships with in the selected study area.

The MS Access report presents the following information for review by Fish and Wildlife biologist: 1) Location, including USGS Quadrangle, Legal, and Ownership reported in the

database and derived from GIS overlay. 2) Percent ownership by federal, state, industrial and non-industrial private timberland

categories 3) Evaluation Status of territory as defined by Gordon within the state dataset 4) Reproductive Status as calculated from Gordon’s data using the SQL processing

sub-system previously described


Mapping products are generated using the data driven capabilities of the Map Book Extension provided with the ArcObjects Developers Kit installation. Each study area within the Northern Interior portion of California (within the area of jurisdiction by the US Fish and Wildlife Yreka office) has a discrete suite of map products at specific scales covering the full study area that are re-produced as an automated map job once data has been re-loaded into and processed through the system. The Map Book extension uses a polygon feature class with features created to define the extent of a specific document. Map specific data driven attributes are calculated with a range of informational values, which area then used to dynamically update a suite of text controls in the map documents. Entire map sets are produced in a single operation using data derived using the various sub-systems in the application.


The new geoprocessing framework provided by ArcGIS 9.0 and the SQL processing capabilities in MS Access have allowed us to develop a system that is fully integrated and automated. Processing time is now on the order of minutes, where previously the work described above would have taken a day or more to complete. All functional sub-systems are implemented within a single MS Access Personal Geodatabase, providing significant efficiencies in system management and portability.


Telemetry Data Analysis The second part of this project has been a research exercise in landscape utilization by local owl populations. The data available for analysis is approximately 4 years of telemetry data developed by NCASI (the National Council for Air and Stream Improvement) in collaboration with local federal and private timberland managers. Fish and Wildlife and land management biologists agreed to cooperatively investigate the relationships of landscape utilization relative to slope position of telemetry points within the study area at various scales. The principle question was “at any given scale, what proportion of telemetry points are found within a given slope range relative to valley bottoms and ridges, and what does that tell us about the requirements of habitat/vegetation management in these areas?” and, in general, “what areas of the landscape are being used when?” The first product we developed was a time series dataset development derived from the telemetry data. A SQL sub-system was used to generate a query recordset object sequenced by individual owl (male or female) by date and time. Additional attributes were calculated into the data to define seasonal values. The sequenced recordset was then processed into time step path features connecting the telemetry locations. These data were used to produce seasonal animations used to visualize the movement of areas of activity over time at approximately a 1:18.000 scale.


The product was delivered as an MS PowerPoint file that allowed the research biologist to move forward or backward in season time steps in their review of the data at their desktops without the need for GIS software on their system. Another time step specific product we developed was a series of ArcScene animations that were used to visualize the discrete movement of the area of activity across an area of interest from record to record within time sequenced and synchronized recordsets. To accomplish this, we again designed as SQL sub-system to calculate query recordset objects sequenced by individual owl by time and date, but this time constrained the recordsets to a subset of telemetry records with common dates between individuals occupying the same nesting territory. These were again processed into time step path features, which were used as the path source to create the ArcScene animations. Multipatch features were used as the animation features to facilitate quality visualizations. This proved to be an endless source of interest to and discussion between biologists as the animation looped continuously. The value of this tool proved to be the types of questions it raised with the biologists relative to the different patterns of movement observed between male and female birds. Each individual displayed clear preferences for landscape types and areas within the extent of their range of activity. The area below is a 7th field watershed within the study area using the data for a single pair of birds.


The next exercise was to define the various scales and principles to be used in the slope position analysis. Standard hydrologic catalogue units are divided ultimately into 7th field watershed units ranging in size from 5,000 to 14,000 acres within the study area for this project. These units were considered too small a scale for an evaluation of slope position utilization at sufficient resolution. This does however represents a scale that can be used to define patterns given the watersheds proximity to major rivers and vegetation gradients in the study area, and as such was of interest, and the smallest scale at which we processed results. For this scale, we derived the high and low elevation values for each 7th field watershed and for each telemetry point, and calculated a slope position value as follows: Slope Position (%) = ((Telemetry Elevation – Low Elevation)/(High Elevation – Low Elevation)) * 100 The scale of most interest to researchers was at the drainage level. In other words, “what area of the landscape within discrete drainages gets the most use?” To investigate this, we employed a range of hydrologic functions to first define the scale of analysis, then to develop an array of hydrologically defined features at that scale with which to conduct that analysis. To define this scale, we ran an initial series of catchment tessellations of the study area at various scales defined by flow accumulation threshold values. Our goal in this exercise was to determine which scale of tessellation resulted in catchment and synthetic stream network features that most closely resembled the mapped routed hydrography developed and managed by the local national forest, as this represented the most comprehensive and widely accepted regional hydrography in the area of interest, and the primary dataset we would be using in modeling runs. USGS 10 meter DEMS were mosaiced, reconditioned by burning the forest hydrography into the DEM using the ArcHydro Agree methodology, and processed into catchment features at flow accumulation thresholds of 1500, 2000, 2500, and 3000 upslope cells. Local variation in the results was considerable, with some areas over and other areas under tessellated. We determined that the 2500 cell threshold tessellation best represented the density of streams mapped in the routed hydrography. We investigated two calculation of slope position using these features. The first was defined relative to the high and low elevation in the catchment polygon, as slope position at the 7th field watershed scale had been calculated. The second was calculated relative to the elevation of the nearest point along the nearest hydrographic feature within the catchment containing the telemetry point. This involved overlaying telemetry points and hydrography with catchments, running a nearest feature procedure for multiple near levels, calculating to the 25 nearest hydrographic features, then developing a SQL sub-system that would process a final results table out of the various primary and derived datasets. Using these data, slope position is calculated as follows: Slope Position (%) = ((Telemetry Elevation – StrmPt Elevation)/(High Elevation – StrmPt Elevation)) * 100


Both of these calculations provided the biologists insight into landscape utilization, but did not fully answer the primary questions, and did not adequately define a slope position calculation relative to functional landscape features. Catchments represent sub-divisions of the landscape used for hydrologic modeling, and do not truly represent drainage units. To address this issue, we needed to process catchment and hydrographic features into drainage specific features. We then needed to define a methodology to define the location of each telemetry point relative to ridge lines as well as valley bottoms. Catchment features were overlain with route hydrography and assigned the LLID value for the primary stream feature within the catchment. Having re-conditioned the source DEM, the forest hydrography still did not fully conform to the catchment delineations, any given catchment might overlay with several hydrographic features (due to the stair-step catchment delineations). We developed a SQL routine to identify the primary LLID associated with each catchment, which was accurate to 98% for bridging, intermediate and terminal catchments alike. The final 2% were identified and edited manually. Once completed, we then dissolved catchments into features representing discrete drainages as defined by the forest routed hydrography. The following screen shot displays results of catchment tessellation and drainage feature processing relative to 7th field watersheds.


With drainage features defined, we next developed the methodology for defining ridge locations relative to telemetry point locations. Each drainage area boundary represents divides which may or may not necessarily represent a ridge. These features can often represent divides on relatively undefined landscape, and are ultimately dependant on the scale of hydrologic tessellation. There are also differences in interpretation between biologists that needed to be considered. Having reviewed and agreed on the scale dependencies, we now needed a practical method to define the ridge / telemetry / valley bottom relationships. For the valley bottom associations we again ran a nearest feature procedure this time for multiple levels of nearness to hydrographic features overlaid with drainage polygons and derived the low elevation as the elevation at the nearest point on the nearest routed stream feature. To define the high elevation point, we decided to develop vector features following the fall line upslope to the point of intersection with the drainage divide. To accomplish this, we first inverted the DEM (value * -1), calculated flow direction and flow accumulation surfaces, then ran a least cost path procedure for each telemetry point, using the upslope flow accumulation grid as the cost surface and the upslope flow direction grid as the back link surface. The resulting least cost path vector features were then routed, intersected with drainage divides, and processed through a SQL sub-system into a resultset that returned the elevation of the first point of intersection along the routed cost path with the drainage divide.


An issue we encountered at this point was that the least cost path features would often run upslope parallel to drainage divides for a considerable distance before actually intersecting those features. This appears to be a consequence of the standard D8 flow direction algorithm, and represented a potential bias in the analysis that would force the telemetry point slope position calculation to a lower value. To resolve this issue, we developed SQL program to calculate events along the least cost path routes at specific intervals, then ran a near function from those events to the drainage divide. We then processed those results through a SQL sub-system to identify and use the elevation for the first point along the least cost path route that returned a distance to drainage divide value below a specified threshold. Slope position for these data was calculated using the following expression: Slope Position (%) = ((Telemetry Elev – StrmPt Elevation)/(EvtCostpath Elev– StrmPt Elev)) * 100

The final step in developing data for evaluation was to develop a slope position value using the slope position aml developed by David Hatfield (US Forest Service, Region 6). This aml runs a grid based procedure using hydrologic functions on elevation and inverse elevation surfaces that define ridges and valley bottoms on the basis of flow accumulation thresholds. We looked at two sets of thresholds values to define the slope position surface, Valley bottom = 2500 cells and Ridge tops = 2500 cells, and


Valley bottom = 2500 cells and Ridge tops = 5000 cells. We selected the valley bottom threshold so that the two analysis methodologies could be directly compared. The ridge top thresholds were selected to evaluate the densification of ridges at those thresholds and its effect on the calculation of slope position for each telemetry point (see below).


Final results of the various calculations of slope position at the various scales explored were processed through a SQL sub-system that calculates the proportion of points in the sample that are within a defined range of slope position values. Results of these calculations are fed into a suite of chart control objects in both an MS Access form and its associated printable reports. The processing sub-system, form, and report objects are interactively integrated such that data can be dynamically binned, displayed and printed in discrete range classifications and for a select 7th field watershed or “All” 7th field watersheds within the study area, providing a flexible exploratory tool for the research biologists conducting this study.


The results consistently show a strong association with slope position defined at the various scales we have investigated. Researches consider this to be one of the most important variables under review at this time. Results of these analyses are currently being reviewed by Fish and Wildlife, land management and consulting biologists involved with the project. Final results will be included in a statistical analysis of a number of variables used to explore the question of landscape utilization, and will be integrated into timber management planning guidelines in the region. Conclusions: The comprehensive integration of GIS and relational database technologies provides a powerful framework for the management, modeling and analysis of biological data. The implementation of the geoprocessing framework with the release ArcGIS 9.0 increases the capabilities and efficiencies achieved through these integrated systems by orders of magnitude. We consider this to be the essential framework for system development, and have implemented the design principles presented in this paper globally throughout the full range of projects in our organization. About the author: Dave LaPlante is a GIS analyst and database developer working with Resource Management and the Northern California Resource Center in Ft. Jones, CA, and has been developing analytical systems for natural resource assessment for the last 5 years. These include a suite of applications for Road Inventory and Assessment, Stream Inventory and Analysis, Sediment Production Analysis, and various biological data management systems, such as those presented here. He is also actively involved in promoting the use and understanding of GIS and database technologies throughout the region as coordinator of the Siskiyou Area ArcGIS User’s Group (SAGU). Resource Management is a natural resource consulting firm. The Northern California Resource Center is a non-profit ecosystem training and assessment entity Both organizations develop and manage projects throughout Northern California and Southern Oregon, and collaborate with a range of clients including the federal agencies (USFS, USFWS, BLM, NRCS), state agencies (CDFG, CDFFP, SWRCB), the Sierra Economic Development District, various Resource Conservation Districts and watershed organizations, as well as private resource management entities throughout the region. He can be contacted at [email protected] (530) 468-2888

GIS Data Management Systems for Territorial Raptors · edit review no_sptl dup_sptl fk12,i13 own_class fk12,i13 owner pvtown_other fk4,i5 state fk4,i5 county usfs_r usfs_f usfs_d

Documents