Update #6
THIS UPDATE PRODUCES IDB VERSION 2.0
Updates TO Version 1.4-1 of the Integrated Database (IDB)
PLOT_ID=2599 in western Oregon FIAWO was dropped because it was out of inventory area
Note: For more information about a column or table that is mentioned below, see the specific column
definitions (using the database form) or the user guide.
New or updated technical documents:
TPA_TPH_Calculations.doc (updated)
Area_Expansion_R6_BLMWO.doc (updated)
Stocking and stocking equations.doc (updated)
Table of differences between snags and mortality sampling.doc
DEAD TREE DOCUMENTATION.doc
Mortality_USING THE TREE_MORT TABLE.doc
Down Wood Inventories in the IDB.doc
Down Wood Calculation_Methods_Ecological Indicators paper.pdf
Down Wood Decay Class codes.doc
Down Wood Paper on the Line Intersect method.pdf
Volume equations.doc (updated)
1.
DataSource: ALL except FIAWW
Table:
DOWN_WOOD table was added to the IDB
Issue:
Information on down wood was added to the IDB for all inventories except FIAWW.
Down wood was collected on either transects or rectangular plots nested within the
boundaries of the standard forest inventory field plot.
This data can be linked directly to the COND table and can be used along with all other
information collected on the plot for a complete stand level analysis. Choose
CWD_sampled=1 from the COND table to select the plots and conditions where down
wood was sampled.
2.
DataSource: ALL
Table:
TREE_DEAD table was added to the IDB
Issue:
Information on all standing dead trees (snags) was added to the IDB for all inventories.
This data can be linked directly to the COND table and can be used along with all other
information collected on the condition. This table is different from the TREE_MORT
table in that all dead trees in the TREE_DEAD table are standing, which is not
necessarily the case for recent mortality data found in the TREE_MORT table
(these can be standing, down, or removed from the site). Choose Snags_sampled_YN=Y
from the COND table to select the plots and conditions where snags were measured.
3.
DataSource: FIAEO
Table:
All
4.
Issue:
Additional data on juniper forest land was added to the database. This data was collected
in a special inventory of juniper forest land in eastern Oregon on state and private lands.
Earlier versions of the IDB had place holder records representing the juniper plot
location. These records were replaced with plot, condition, and tree data that were
collected at these locations.
DataSource:
Columns:
Table:
Issue:
ALL
Live_trees_sampled_YN, Mortality_sampled_YN, Snags_sampled_YN
COND
These columns were added to the COND table to identify which conditions and plots
were intentionally sampled for live trees, mortality trees, and standing dead trees (snags).
Not all plots were sampled for all components for a variety of reasons.
All classes of forest land, reserved land, new plots, and access denied plots may not have
been sampled by design. Please select the column and choose ‘Y’ to get the correct set of
Plots and conditions for an analysis or summary of data.
Mortality_sampled_YN: Indicates if the condition was sampled for mortality.
The western Washington FIA inventory intensification was NOT sampled for mortality
(the intensification was a doubling of the number of field plots normally collected on the
FIA sample grid; because these plots were all new, mortality was not recorded or
estimated on this set of plots) .
FIA Oregon inventory plots that were access denied had NO mortality estimated on them.
Note that the FIA California inventory DID estimate mortality on these conditions in the
office by partioning the TPA of trees measured in the past.
Also, some FIA GLC’s were not sampled for mortality (40’s in Washington, some 40’s in
California). Only condition class #1 was sampled for mortality in all FIA inventories, on
remeasured subplots. Reserved lands were not always sampled for tree data.
FIA subplots labeled as ‘N’ were not sampled for mortality in FIACA, FIAEO, FIAWO.
Choosing Mortality_sampled_YN=Y will insure that the correct set of plots are selected
for an analysis of mortality.
5.
DataSource:
Column:
Table:
Issue:
ALL
CWD_sampled
COND
This column was added to the COND table to identify which conditions and plots
were intentionally sampled for down wood in each inventory.
All classes of forest land, reserved land, new plots, and access denied plots may not have
been sampled by design. Please select the column and choose “1” to get the correct set
of plots and conditions for an analysis or summary of data.
6.
DataSource:
Columns:
Table:
Problem:
FIA
SURVEY_UNIT_FIA
PLOT
Benton county (3) in Oregon is listed in the wrong survey unit in the 1995 Field Manual.
It says it’s in the Northwest unit, when really it is in West Central unit.
The wrong survey unit code was found for the following states and counties:
OREGON :
CTY=69 SHOULD BE SURVEY_UNIT_FIA=3 (WAS 4)
CTY=47 SHOULD BE SURVEY_UNIT_FIA=0 (WAS 1) (1 plot)
WASHINGTON
CTY=23 SHOULD BE SURVEY_UNIT_FIA=9 (WAS 8) (42 plots)
CTY=17 SHOULD BE SURVEY_UNIT_FIA=8 (WAS 9) (14 plots)
CALIFORNIA
CTY=91 SHOULD BE SURVEY_UNIT_FIA=3 (WAS 1) (1 plot)
Fix:
The PLOT table was updated with the right survey unit codes.
DataSource:
Columns:
Table:
Problem:
R6
SITE_INDEX_FIA, MAI,GLC
COND
An error was found in the site index equation for LARCH which affects Region 6 plots.
This caused some conditions with a larch site index to be classified incorrectly as a low
site class.
Fix:
Site index was recalculated, MAI was recalculated, and the GLC code was reassessed to
see if it is timberland or other forestland (these determinations are based on MAI).
Some of the GLC codes have changed from GLC 49 (low site) to GLC 20 (timberland).
DataSource:
Columns:
Table:
Problem:
Fix:
many
LATITUDE_FUZZ, LONGITUDE_FUZZ
PLOT_COORDINATES
Some plot coordinates were incorrect for Region 5 plots.
All fuzzed coordinates were checked and updated if necessary.
7.
8.
9.
DataSource:
Columns:
Table:
Problem:
Fix:
All
FVS_VARIANT
PLOT
FVS variant classifications were not available for FIA plots in the past.
A new GIS cover was provided by Region 6 staff to help assign a variant to every FIA
plot for the first time; and to update the existing variant codes on National Forest and
BLM plots.
New or updated FVS variant codes were added to all PLOT records.
Note that the ancillary table called IDB_CVS_VARIANT has been removed from the
IDB because the PLOT table now contains the official variant code.
10. DataSource:
Columns:
Table:
Problem:
Fix:
R6
BA_FT2_AC , BA_M2_HA
TREE_LIVE
Trees with a PLOT_SIZE of 1 or 2 had incorrect values calculated.
The BA_FT2, BA_M2, TPA and TPH were correct, so the BA_FT2_AC and
BA_M2_HA were recalculated by multiplying the tree basal area by TPA or TPH.
11. DataSource:
Columns:
Table:
Problem:
FIA inventories
DATA_SOURCE
TREE_LIVE
Some tree records that had an R6 Data_Source on them, when they should have had and
FIA data source (this affects COND_ID’s = 133,1450,3347,18420,30791)
Recoded DATA_SOURCE as follows:
Fix:
COND_ID =133: The DATA_SOURCE was changed from R6 to FIAWO
COND_ID =1450: The DATA_SOURCE was changed from R6 to FIAWO
COND_ID =3347: The DATA_SOURCE was changed from R6 to FIAEO
COND_ID =18420: The DATA_SOURCE was changed from R6 to FIACA
COND_ID =30791: The DATA_SOURCE was changed from R6 to FIAEW
12. DataSource:
Column:
Table:
Issue:
ALL
Mortality_sampled_YN (new column)
SUBPLOT
Indicates if the subplot was sampled for mortality; Important because some FIA subplots
were not sampled. These are mainly N (NEW) subplots in all FIA inventories except
eastern Washington -- and the non-intensified sample of western Washington, and access
denied subplots (‘P’ ) in Oregon. Only remeasured subplots were sampled for mortality.
13. DataSource:
Column:
Table:
Issue:
FIA inventories
FIELD_SUBPLOT_NUMBER (this is a NEW column)
SUBPLOT
We thought it would be helpful to include the original subplot code from the FIA
databases in the IDB because the code itself contains meaningful information. For NFS
and BLM inventories, this column is the same as SUBPLOT_NUM.
Added this column to the SUBPLOT table
Fix:
14. DataSource:
Column:
Table:
Issue:
Fix:
15. DataSource:
Column:
Table:
Issue:
Fix:
A few
SAWTIMBER_YN
TREE_LIVE
Some hardwoods that were less than 11” DBH (437 trees) had the wrong
sawtimber code.
Changed the SAWTIMBER_YN code from Y to N on these hardwood trees
R5
SUBPLOT_COUNT_PLOT
PLOT
The count on some plot records was incorrect.
Updated the field with the correct count of subplots installed on the plot.
16. DataSource:
Columns:
Table:
Issue:
Fix:
17. DataSource:
Columns:
Table:
Issue:
Fix:
ALL
GRID_WT_FAC, GRID_WT_MORT_FAC (new columns)
COND
Data were needed to provide a weight for each condition that accounts for the many
sampling grid intensities that occur in the IDB.
Added 2 columns that provide a weight relative to the 3.4 mile grid, which has a
weight of 1. The separate grid weight for mortality estimates was needed because
recent mortality was not sampled on all plots across the inventories. These columns
were used to create sampling weights (below) which, in turn, should be used to
calculate weighted averages of per-unit area estimates, if desired.
ALL
SAMPLING_WT_FAC, SAMPLING_WT_MORT_FAC (new columns)
COND
Data were needed to provide a weight for each condition that accounts for both
sampling grid intensities and the size or area of the condition class on the plot.
These columns were calculated and added to the COND table.
SAMPLING_WT_FAC = GRID_WT_FAC*COND_WT. These weights should be
used to calculate weighted averages of per-unit area estimates, if desired. A per-acre
column should be summed up for a condition, followed by calculating a weighted
mean across conditions.
18. DataSource:
Column:
Table:
Issue:
R6
ACRES, ACRES_VOL
COND
An error was found in the number of acres assigned to wilderness areas in the Umatilla
NF. After correcting the error, we needed to make an adjustment to the total acres within
each county. Our goal has always been to match the total acres per county in the IDB
with the county land area published in the 1990 Bureau of Census reports (because
inventory data were collected in the 90’s). At times it was difficult to mesh the national
forest acres with state and private land area to produce the correct estimate of land area
by county (because the acres came from 2 different inventories). To try to improve these
estimates, we decided to recalculate the acres per plot for Region 6 using land area
estimates published on the national website that tracks national forest land area (both
proclaimed and administered) for each county in the country. This method allows the
total land area to match Bureau of Census figures for a given county. The stratification
uses the amount of Proclaimed acres of total National Forest land within each county.
The number of NF plots counted in each county are divided into the NF land area within
the county (irregardless of specific National Forest) to derive the acres per plot. Users can
then summarize the amount of acres by the Administered National Forest code, which is
the most common way to analyze and report NFS data. The result was a small
adjustment to the ACRES and ACRES_VOL columns for Region 6.
19. DataSource:
Columns:
Table:
Problem:
Fix:
All
QMD_HWD, QMD_SWD, QMD_TOT
COND
Column names were not explicit enough.
Changed the column names to QMD_HWD_CM, QMD_SWD_CM, QMD_TOT_CM, to
more accurately reflect the units of the data (centimeters).
20. DataSource:
Column:
Table:
Issue:
Fix:
FIAWO
OWNER, OWN_GROUP
COND
Some owner codes were incorrect
All owner codes were checked and updated for western Oregon FIA
21. DataSource:
Column:
Table:
Issue:
Fix:
22. DataSource:
Column:
Table:
Issue:
Fix:
ALL
ID
All tables
In the IDB, each table had a column called “ID” which uniquely identified every row in
the table. Repeating the same column name in multiple tables has the potential to cause
confusion, especially when these columns are key fields and used to join other tables.
To improve clarity and reduce the potential for mis-joins, the ID column name in every
table was changed to include the name of the table. For example, in the COND table this
column changed from ID to COND_ID and in the TREE_LIVE table it is now called
TREE_LIVE_ID.
FIACA
SITE_INDEX_FIA
COND
When the SITE_SPECIES=299 and SITE_CLASS_DUNNINGS =0 the programs used a
mixed conifer site index equation. In FIA inventories, each condition had a mixed
conifer site indicator in the original database. The mixed conifer site index equation
should only be used when the crews identified the plot as a mixed conifer site.
Some conditions with a SITE_SPECIES=299 and SITE_CLASS_DUNNINGS =0 were
not on a mixed conifer site, therefore the site index was calculated from the wrong
equation.
The SITE_INDEX_FIA was recalculated for this set of plots where SITE_SPECIES=299
and SITE_CLASS_DUNNINGS =0 and the plot was not a mixed conifer site, using a
more appropriate site index equation. The SITE_SPECIES code was changed to the
predominant softwood species.
23. Datasource: FIAEO
New data records : New data became available for the eastern Oregon FIA inventory. Specifically,
data from a special inventory of juniper forest lands was added to the IDB. This data includes
measurements of trees and other attributes on juniper forest land. The new records replace existing records
which were placeholders in the PLOT and COND table. The plots have the same PLOT_ID and
coordinates as before, and now include tree, down wood, and snag data as well.
24. Datasource: RMRS
A new datasource was added to the IDB. This includes data on plots collected by the Rocky Mountain
Research station in National Forests that crossed over into Washington and California. The RMRS collects
data for Region 1 (the Kaniksu in eastern Washington) and Region 4 (the Toiyabe in eastern California)
which are now included in the IDB. These sections complete the inventory assessment for these states.
25. DataSource:
Column:
Table:
Issue:
Fix:
26. DataSource:
Column:
Table:
Issue:
Fix:
R6, FIAEW
ELEV_FT and ELEV_M
PLOT
Elevation for these data sources was off by a factor of 10.
The existing elevation was multiplied by 10 and updated in the database.
ALL
HWD_FORM_CLASS
TREE_LIVE
Some of these codes were the original codes from the original inventory databases.
The form class codes were recoded with a 1 or a 2, according to IDB documentation.
Only hardwood trees >=12.5cm should have a code in this column. The RMRS Data
Source did not collect this data, all records are null.
27. DataSource:
Column:
Table:
Issue:
Fix:
28. DataSource:
Column:
Table:
Issue:
Fix:
29. DataSource:
Column:
Table:
Issue:
Fix:
30. DataSource:
Column:
Table:
Issue:
Fix:
R6, Southwest Oregon plots
PLANT_ASSOC_CODE
SUBPLOT and IDB_PLANT_ASSOCIATIONS
The plant association codes (PA) for southwest Oregon were based on a system
developed by Tom Atzet. All other plant associations were based on a system by Fred
Hall. The codes are very similar but have different association names. When linking the
SUBPLOT table to the IDB_PLANT_ASSOCIATIONS crosswalk table, the plots in
southwest Oregon would be linked to a slightly incorrect association name.
The PLANT_ASSOC_CODE was updated by inserting an ‘X’ before each code on
southwest Oregon subplots (R6 data) to distinguish them as codes based on Atzet’s
system. In addition, these “X” codes were added to the IDB_PLANT_ASSOCIATIONS
table along with a new plant association name.
ALL
TREE_FLAG (new column)
TREE_LIVE, TREE_MORT
During early compilation and review of the data in the IDB, certain trees were identified
as having an unreasonable height or diameter. In some cases the height was very large or
very small. In other cases the height to diameter ratio appeared to be unreasonable.
Heights were also compared to a calculated height using the FVS height estimation
equations. A tree was flagged if the tree's height was more than twice the calculated
height or less than half the calculated height. Before Version 2.0, these trees were not
included in the IDB database.
It was decided that all trees tallied on a plot should be included, regardless of the
apparent quality of the data. We flagged these trees, allowing users the option of
choosing to keep these trees in their analyses or excluding them. We calculated new
heights on most trees followed by stocking, volume and biomass estimates. These trees
can be identified by a TREE_FLAG>0 and HT_EST_METHOD=2 or 5. The result is
that Version 2 contains some new tree records in the TREE_LIVE and TREE_MORT
tables. In addition, the new trees on a condition have changed the forest types and stand
size on some plots. Please refer to the last page for more information on the addition of
trees to the IDB and the impact that this had.
R6
TREE_SPECIES
TREE_LIVE
There were 3 tree records with a tree species code of 92 (Brewer spruce) in Washington.
Trees were on COND_ID= 33206. We believe this species does not grow in Washington.
The species code was changed to Unknown Softwood (code 298)
R5
TREE_SPECIES
TREE_LIVE
There were tree records with a tree species code of 211 (Redwood) in Lassen, Madera,
and Trinity counties in California. Trees were on COND_ID= 23187, 24219, 24411,
21577, 23718. We believe this species does not grow in this area.
The species code was changed to Unknown Softwood (code 298)
Other---•
A folder has been added to the CD containing PDF files of all field manuals used in the
inventories in the IDB.
•
Some users have requested more identifiers on records of some tables. We added STATE, CNTY,
PLOT, FOREST_OR_BLM_DISTRICT, and COND_CLASS to the end of each record in the
COND, TREE_LIVE, TREE_DEAD, and DOWN_WOOD tables. These columns are meant to
serve as a reference while using the data – we recommend that users still link tables with the
appropriate ‘ID’ column. One reason we created the ‘ID’ column was to make linking simple and
easy. In most cases, you only need to use one column as a key field for linking.
If you choose to link based on added STATE, CNTY, PLOT, FOREST_OR_BLM_DISTRICT,
and/or COND_CLASS, remember that you need all of them to make the proper linkage.
A query was added that combines data from the PLOT, COND, and TREE_LIVE tables, and creates
one table with all this information on one record. The query is a ‘make-table’ query and takes a
significant amount of time to run. This new table may make the inventory data appear closer to the
format of some of the original FIA databases and files, that some users may be familiar with.
A series of queries were created that provide an estimate of weighted means for down wood variables.
A SAS program was also created and included, which provides another method of estimating weighted
means.
Additional information about trees that were previously excluded from the IDB
During the early compilation and review of the data in the IDB, certain trees were identified as having an
unreasonable height or diameter. In some cases the height was very large or very small. In other cases the
height to diameter ratio appeared to be unreasonable. Heights were compared to a calculated height using
the FVS height estimation equations, and if they were more than twice the height or less than half the
height of the FVS estimate, the tree was flagged and removed from the IDB. The reason they were
removed was because the IDB staff could not be certain if the diameter was in error or the height was in
error. In most cases, the measurements on these trees caused very strange (and unusable) volume and
biomass estimates. Some estimates were very high, extremely low, negative, etc. because the volume
equations were designed with normally formed trees and did not perform well on trees with an unusual dbh
to height relationship. IDB staff felt these trees would reduce the overall quality of the data if they were
included in the database.
Recently, however, after another review of the data, a decision was made to establish the DBH as the
correct measurement. This would imply that the height was in error on the set of trees that were removed
from the database. For version 2.0, we added this set of trees back into the IDB and calculated a new
height using the FVS equations. All subsequent tree calculations were based on the original diameter and
newly calculated height. Volume, biomass, stocking, etc were calculated on these trees. In summary, all
trees originally tallied on a plot are now included in the IDB, with the assumption that the DBH is the most
likely measurement to be correct on trees that have a questionable DBH to height relationship.
We flagged these trees, allowing users the option of choosing to keep them in their analyses or to exclude
them. The new column called TREE_FLAG contains the following codes:
1
2
3
5
6
9
Unreasonable height: height was less than ½ of the FVS calculated height for the same DBH
Unreasonable height: the height to DBH ratio was less than 1.5
Unreasonable height: height was more than twice the FVS calculated height for the same DBH
Unreasonable height: height appears to be too small (short) for the size of the tree
Height was missing from original data
Unreasonable height: height produced unrealistic results for volume and biomass, new equation needed
All other trees that IDB staff did not question have a TREE_FLAG value of null. Basically, when
TREE_FLAG >0 we questioned the DBH to height relationship and made a decision to recalculate the
height along with all other relevant tree variables.
If a new height was calculated, the HT_EST_METHOD =2 or 5. In some cases a new height was not
calculated because the growth habits of the particular species often produce trees of irregular form, and the
volume equation available was based on these characteristics.
The result is that Version 2 contains some new tree records in the TREE_LIVE and TREE_MORT tables.
Trees were added across many plots, and in most cases this had no impact on condition level calculations.
However the forest type, stand size and stand age were recalculated and in some cases these classification
variables changed based on the additional trees. The most common change was from a non-stocked forest
type to a stocked type.
In summary, all trees originally tallied on a plot are now included in the IDB, with the assumption that the
DBH is the most likely measurement to be correct on trees that have a questionable DBH to height
relationship. Users may not agree with this assumption and may prefer to use the original height measured
or estimated for the tree. These original heights were retained in a separate table called
XTRA_REF_ORIGINAL_HEIGHTS_ON_QUESTIONABLE_TREES. This table has tree-level
identifiers as well as species, dbh, and the original height. Users may want to evaluate the quality of the
data using their own criteria, and choose to use the original height instead of the height calculated by IDB
staff. This table can be linked to the TREE_LIVE table via the TREE_LIVE_ID or to the TREE_MORT
table via the TREE_MORT_ID if users want to add the original height back onto the tree record.