Inventory data ingestion process

Adriana Tovar

2024-07-17

library(forestables)
#> Loading required package: data.table
#> Loading required package: dtplyr

Introduction

This vignette documents the most important processes in handling official data sources from the national forest inventories of France, the United States, and Spain.

IFN (SPAIN): data processing

The National Forest Inventory (IFN) is a project that provides national-level information on forests and their evolution from both a taxonomic and ecological perspective. With a periodicity of at least ten years, it has been ongoing for over fifty years.

The Second National Forest Inventory (IFN2) was conducted by the Forest Inventory Service between 1986 and 1996. It consisted of permanent forest plots marked using a metal rod buried at the center of each plot, allowing for the periodic measurement of each tree and the comparison of their physical characteristics and condition.

The Third cycle of the National Forest Inventory (IFN3) was carried out between 1997 and 2007. While it followed the methodology established in IFN2, with a focus on continuous monitoring of the same forest plots, it introduced changes by considering forest ecosystems for the first time in Spain. IFN3 aimed to provide insights into the state and evolution of forests at the provincial level through over 100 indicators.

The Fourth cycle of the National Forest Inventory (IFN4) started in 2008. While IFN4 continues the methodology of its predecessors, it incorporates some changes and innovations. Among the main novelties of this fourth cycle (IFN4) is the use of the Forest Map of Spain at a scale of 1:25000 (MFE) as the cartographic base.

Data acquisition:

• IFN2 : Field data bases (DBF files):

  Data is organized in provinces and there are thematic DBF files with similar names for all provinces.

  https://www.miteco.gob.es/es/biodiversidad/servicios/banco-datos-naturaleza/informacion-disponible/ifn2_descargas.html

• IFN 3: Field data bases (Access 2007 *.accdb)

  Data is organized in provinces and each province one has an access file. Within the access files there are thematic sheets with similar names.

  https://www.miteco.gob.es/es/biodiversidad/servicios/banco-datos-naturaleza/informacion-disponible/ifn3_bbdd_descargas_htm.html

• IFN4 : Field data bases (Access 2007 *.accdb)

  Data is organized in provinces/ Autonomous community with an internal organization similar to IFN3.

  https://www.miteco.gob.es/es/biodiversidad/temas/inventarios-nacionales/inventario-forestal-nacional/cuarto_inventario.html

Individual functions for IFN tables

Variables have been harmonized between inventories.
Some variables may appear on multiple tables:

Harmonized name id_unique_code	Original name IFN2 ES_ESTADILLO_ PROVINCIA	Original name IFN3/ 4 ES_Estadillo_Provincia	Description Code to identify plot
province_code	PROVINCIA	Provincia	Territorial division
plot	ESTADILLO	Estadillo	Plot ID
class	CLA	Cla	Code that indicates when the plot was surveyed.
subclass	Subclase	Subclase	Code that determines the type of plot based on the re-location of the metal rod (“rejon”).

IFN TREE TABLE PROCESSING

Information is obtained from PCMAYORES table of IFN3-4 and PIESMA of IFN2
Tree condition: In the IFN trees must have a diameter (DBH) greater than 7.5 cm to be sampled.

Harmonized name	Original name IFN2	Original name IFN3/ 4	Description
height (m)	ALTURA (m)	HT (m)	Height of tree
sp_code	ESPECIE	Especie	Code of species at IFN
tree_ifn2	-	OrdenIf2	Order number (ID) to trace individuals between inventories
tree_ifn3/tree_ifn4	-	OrdenIf3/OrdenIf4	Order number (ID) to trace individuals between inventories
tree_id	ARBOL	nArbol	number of trees
dbh (cm)	DIAMETRO1 (mm)	Dn1 (mm)	diameter in one direction
dbh (cm)	DIAMETRO2 (mm)	Dn2 (mm)	diameter in perpendicular direction
cubing_form	FORMA	Forma	Measurement method of the foot (see ifn3 documentation)
wood_quality	CALIDAD	Calidad	Quality of the measured foot ( 6 = dead) (see ifn3 documentation)

The mean of diameters is calculated to obtain the variable dbh (diameter/DBH), and it is converted to cm.

  dbh = ((Dn1 + Dn2) / 2) * 0.1, # From mm to cm

We add a new variable called density_factor, corresponding to the Expansion Factor per hectare of the unit values of different dendrometric parameters of the measured stem. IFN tree sampling follows a variable-radius plot sampling, where smaller trees are measured at smaller radii of the plot, while larger trees are measured within the larger radii as follows.

DBH interval	Radius sub-plot	Density value
[0, 12.5)	5	127.3239546
[12.5, 22.5)	10	31.83098865
[22.5, 42.5)	15	14.14710607
[42.5, 999)	25	5.092958185

We also add a new variable: sp_name , with the scientific name of the species that corresponds to the sp_code.

The final variables provided are id_unique_code, province_code, class, subclass, plot, sp_code, sp_name, tree, wood_quality, cubing_form, tree_ifn2, tree_ifn3, tree_ifn4, dbh, height and density_factor.

IFN SHRUB TABLE PROCESSING

Information is obtained from tables PCMatorral of IFN3-4 and MATORR of IFN2

Harmonized name	Original name IFN2	Original name IFN3/ 4	Description
height (cm)	ALTUMED (dm)	Hm (dm)	Mean height
sp_code	ESPECIE	Especie	Species code IFN
cover	FRACCAB	Fcc (%)	Percent cover of the species in the plot

We also add a new variable sp_name, with the scientific name of the species that corresponds to the sp_code.

The final variables provided are id_unique_code, province_code, class, subclass, plot, sp_name, sp_code, height and cover.

IFN REGEN TABLE PROCESSING

Information is obtained from the tables PCRegenera of IFN 3-4 or PIESME of IFN2

Harmonized name	Original name IFN2	Original name IFN3/ 4	Description
height (cm)	ALTUMED (dm)	Hm (dm)	Mean height
sp_code	ESPECIE	Especie	Species code IFN

The regeneration data is sampled differently in the IFN3-4 compared to the IFN2, thus the processing is also different:

IFN2

Original variables are:

Harmonized name	Original name IFN2	Description
-	Numero	Number of trees of a given species with a diameter between 25 and 75 mm in the 5m radius
-	Regena	averaged number of trees of a given species with a diameter below 25 mm in the 5m radius Regena 0 –> n = 0 Regena 1 –> 0<n<5 Regena 2 –> 5<= n<=15 Regena 3 –> n>15
height (cm)	ALTUMED (dm)	Mean height

To harmonize all information and separate each record in a different row some assumptions are made. Additionally, new variables are added (dbh, density_factor and n).

• n ( actual density in trees/ ha ) is calculated as the product of the number of trees and the density_factor.
  For those records with REGENA codes, the mid point of the interval is assumed to be the actual number of species .

Category	Height default	DBH default	Density factor	Numero default	n
DBH below 25 mm	100cm	1 cm	127.3239546	Regena 0 = 0 Regena 1 = 2.5 Regena 2 = 10 Regena 3 = 20	Numero * density_factor
DBH above 25 mm	ALTUMED	5 cm	127.3239546	Numero	Numero * density_factor

IFN3 - 4

Original variables are:

Harmonized name	Original name IFN3/4	Description
-	CatDes	development category of regeneration
-	NumPies	number of trees for category 4
-	Densidad	interval of number of trees for category 1,2 and 3
height (cm)	Hm (dm)	mean height for category 4

Depending on the development category (CatDes), the regeneration densities will be quantified differently by IFN. For development categories 1, 2, and 3, counting the trees in the circular plot of 5 m radius and classifying the density according to the following scale:

Densidad (Density)	Description	NumPies default
1	Sparse: 1 to 4 trees in the plot.	2.5
2	Normal: 5 to 15 trees in the plot.	10
3	Abundant: More than 15 trees in the plot.	20

For development category 4, NumPies contains the number of individuals by species present in the 5-m radius subplot and Hm contains the average total height of each group (in cm).

Additionally, new variables are added and some assumptions are made to uniform the information:

• density_factor: Expansion Factor per hectare which is always 127.3239546

• dbh: Diameter at breast height in cm

• height: Height, for CatDes 1,2 and 3 in cm

• n: Density in individuals per ha calculated as the product of density_factor and NumPies

CatDes 1	Description Trees with height less than 30 cm.	DBH (cm) 0.1	Height (cm) 10	NumPies 2.5	density_factor 127.3239546	n 2.5 * 127.3239546
2	Trees with height between 30 and 130 cm.	0.5	80	10	127.3239546	10 * 127.3239546
3	Trees with height between 30 and 130 cm.	1.5	100	20	127.3239546	20 * 127.3239546
4	Trees with height greater than 130 cm and normal diameter between 2.5 and 7.5 cm. This corresponds to smaller trees (PIESMENORES) in IFN2.	5	Hm	NumPies	127.3239546	NumPies * 127.3239546

The final variables given are id_unique_code, province_code, class, subclass, plot, sp_code, sp_name, dbh, height, n and density_factor

IFN PLOT TABLES PROCESSING

Information is obtained from tables DATEST of IFN2 and PCDatosMap, PCParcelas of IFN3-4

Note that coordinates are only approximate.

Firstly, names are harmonized.

Harmonized name	Original name IFN2	Original name IFN3/ 4	Description
coordx	COORDEX	CoorX	Coordinate longitude
coordy	COORDEY	CoorY	Coordinate latitude
slope (%)	MAXPEND	MaxPend1	Maximum slope (%)
elev (m)	ALTITUD2	-	Elevation above sea
year	ANO	Ano	Year
aspect (sexagesimal degrees)	ORIENTA2	Orienta1	Aspect
huso	Huso * default values	Huso	UTM HUSO

Additionally, two variables are added to indicate coordinate system (coord_sys) and CRS - ESPG code (crs). They are called and crs. We also add a variable indicating version of IFN (version).

IFN2

Coordinates are corrected where letters refer to numbers .

slope is assigned following IFN categories:

Value at IFN	Description at IFN	Default value
1	0-3 % of slope	1.5%
2	3-12 %	7.5%
3	12-20%	16%
4	20-35%	27%
5	More than 35%	40%

The elevation is indicated as the lower limit of the range of 100 in which the actual altitude is found. For example, if it’s between 0 and 100 meters, it’s indicated as 100; if it’s between 100 and 200, it’s indicated as 100, and so on.

Huso (UTM zone) is not recorded in IFN2. Therefore, a default value of 28 is assigned for the Canary Islands, and for the rest of the peninsula, the default value is 30. Similarly, for variable coord_sys default values ar set as follows: ” WGS84” for Canary Islands and “ED50” for the rest of peninsula.

IFN3-4

slope is assigned following IFN categories:

Description at IFN : E20 ESCALE	Default value
[0, 0.6 ]	1.5%
(0.6, 2.4]	7.5%
(2.4,4]	16%
(4-7]	27%
(7-99)	40%

aspect (ORIENTA2) :
It is measured in the field. We have transformed aspect from centesimal degrees to sexagesimal degrees by multiplying by 0.9.

Values of variables coord_sys and huso are assigned following the technical documents of IFN. EPSG code of the coordinate reference system (crs) is obtained from both. When huso is missing (IFN3) , it is assumed that its value is Huso = 30. (This will be corrected in near future).

Final variables are id_unique_plot, country, ca_name_original, province_code, province_name_original, plot, class, subclass, coord_sys, year, version, type, aspect and slope.

FFI (FRANCE): data processing

Since 2005, an adaptable statistical sampling method has been used annually across metropolitan territories in France. This method combines data from five yearly campaigns to generate precise national and regional results. Sampling is optimized by revisiting points every five years, creating a grid system that enhances logistical efficiency.

Data acquisition:

FFI is downloaded from the official portal: https://inventaire-forestier.ign.fr/dataifn/
Version used in the development of this package is 2023-10-11

The data is grouped into 7 files (data tables):

• ARBRE.csv

• BOIS_MORT.csv

• COUVERT.csv

• ECOLOGIE.csv

• FLORE.csv

• HABITAT.csv

• PLACETTE.csv

These CSV files have information for every plot and year.

Additionally, in the downloaded folder you can find important documentation in a PDF format.

• espar-cdref13.csv

• metadonnees.csv

• IGN_DB_documentation_generale_20xx.doc

Version of the first two files are included in the package but note that they were downloaded in October 2023.

Individual functions for FFI tables

Variables that appear in multiple tables:

Harmonized name	Original name FFI	Description
id_unique_code	FR_DEP_IDP
plot	IDP	Plot ID
year	CAMPAGNE	Year
department	DEP	Territorial division

FFI TREE TABLES PROCESSING

Information is obtained from a table (ARBRE.CSV)

In general, only trees deemed censusable (RECENSABLE) are considered in the inventory, with a circumference at 1.30 m equal to or greater than 23.5 cm (7.5 cm in diameter).

The radius of the measurement plot varies according to the diameter of the trees to be measured.

2nd visit

Case of “second visit” points, raw data is also provided on “second visit” locations. The trees and plots concerned are present in the raw data files with those collected on the first visit. The VISIT data present in the placette.csv file allows you to know whether the point is subject to the first or second visit protocol.

From 2010 to 2014, the forest inventory carried out a revisit of the locations seen 5 campaigns previously. This revisit was solely intended to quantify forest harvesting. It was only carried out on inventory plots containing at least one censusable tree during the first visit. On these plots, the cutting (or not) of the trees present on the first visit was mainly noted.

With the 2015 campaign, information is becoming richer on this type of plot. All the trees measured during the first pass are remeasured (or noted as cut or dead) and the new censusable trees present on a 5 m plot are inventoried.

From 2016 on, all the points visited in the field for the first time are the subject of a second visit. In addition to observations and measurements made on the trees, field agents carry out new observations such as those of land cover and use, collect data relating to the evolution of the population (planting, cutting, etc.). The data collected during the 2nd visit are therefore generally less numerous (especially before 2016) and some are not taken (ecological data in particular).

IMPORTANT NOTE:

Check metadonnees.csv and recapitulatif_metadonnes.csv together with the specific documentation of each source file (ARBRE, ECOLOGIE, PLACETTE, COUVERT, FLORE…) to understand if the variable is static of have changed over time, also to see if units (codes ) are the same each year or not. You may need to check on that for example for the meaning of status (veget) or veget5 (revisit), as there have been some adding and changes in the units.

Example:

The same happens with ESPAR. In the package we give as default last version of correspondence between ESPAR, CD_ref and LATIN NAME (espar-cdref13.csv) (OCTOBER 2023). This is based on TAXREF V13 (https://inpn.mnhn.fr/telechargement/referentielEspece/taxref/13.0/menu), as indicated in the french documentation, but could change in the future.

Harmonized name	Original name FFI	Description
sp_code	ESPAR/cd_ref	Species code
tree	A	Tree id
dbh (cm)	C13(m)	Circumference at 130cm height (DBH) If the plot is being revisited it will only contain data after 2015.
height (m)	HTOT (m) | Total height, not recorded in revisited plots.
density_factor	W	Weight/ Expansion Factor per hectare
status/status5	VEGET/VEGET5	Status of the tree in first visit (VEGET) or second visit (VEGET5) See documentation for code.

To transform C13 in m to dbh in cm:

dbh = (C13 / pi) * 100

Final variables are id_unique_plot, plot, dep, year, tree, espar, sp_code, sp_name, status, status5, dbh, height and density_factor.

Note that status5 is status in 2nd visit (after 5 year).

FFI SHRUB TABLES PROCESSING

Information is obtained from a table (FLORE.CSV) that contains information for all kind of species, trees shrub and herbs but he exclusively give information about shrubs. For that we have used R package GIFT (https://cran.r-project.org/web/packages/GIFT/index.html) (See appendix 1) .
Therefore, in the table a variable called growth_form is used to inform about whether it is a tree, a shrub or herbs.

If a species record is not in the GIFT package, automatic classification cannot be done. Therefore, all NA are included in the shrub table ( same happens for regen table).

Harmonized name	Original name FFI	Description
sp_code	CD_REF	Species code (metadata)
cover	ABOND	Percent cover of a species within a plot

ABOND ORIGINAL	ABOND Description	Default value
1	Doubtful presence. Species coverage rate below 5%“.	5 %
2	Assured presence. Species coverage rate below 25%	12.5%
3	Species coverage rate between 25% and 50%	37.5%
4	Species coverage rate between 50% and 75%	62.5%
5	Species coverage rate above 75%“.	87.5 %

Final variables are id_unique_code, plot, dep, year, sp_code, sp_name, cover, height and growth_form.

FFI REGEN TABLES PROCESSING

Table for regeneration is obtained via two ways depending on the year because the protocol changed in 2015 - as it is explained in the IGN documentation. Before 2015, table COUVERT.CSV is used and after FLORE.CSV is used (the same as in shrub table process).

“Starting from the 2015 campaign, the non-measurable stratum no longer undergoes determination of coverage rate through TCL and TCA data from the COUVERT.csv file. Instead, this information is replaced by the ABOND data (from the FLORE.csv file). Consequently, STRATE only has one modality (R = measurable).”

Here we first explain process for COUVERT.CSV.

Harmonized name	Original name FFI	Description
sp_code	ESPAR_C/ CD_REF	Species code
cover	TCA	% cover of a species within a plot
	STRATE	Variable that determines if the species recorded is recensable

We select the records with STRATE = “NR” , meaning non-measurable as TREE, (i.e., dbh< 7,5 cm).

We add a new variable with a default value : dbh = NA ( but maximum dbh present is 7.4 cm).

Final variables are “id_unique_code”, “plot”, “dep”, “year”, “sp_code”, “sp_name”, “cover”, “dbh”, “height”, “growth_form”

Process for FLORE.CSV. (AFTER 2015)

Here same process of SHRUB table process is performed but in the end only those with a growth_form equal to “tree” are selected. Note that this data does not have dbh, therefore, dbh = NA.

If a species record is not in the GIFT package, automatic classification cannot be done. Therefore, all NA are included in the regen table .

Final variables are id_unique_code, plot, dep, year, sp_code, sp_name, cover, dbh, height and growth_form.

FFI PLOT TABLES PROCESSING

Information is obtained from two tables: PLACETTE.csv, that contains basic information about the plot (VIST, COORDINATES), and ECOLOGIE.CSV that contains ecological information. Ecological information is extracted at plot level from the last year with available information because the variables are not recorded every year.

Harmonized name	Original name FFI	Description
visite	VISITE	Indicates whether the point is from the first or second visit to the field
coordx	XL	Coordinate longitude
coordy	YL	Coordinate latitude
aspect (º)	EXPO	Aspect
slope (%)	PENT2	Slope in percentage
lign1_pct	LIGN1	Represents the ratio of the area covered by low woody vegetation (less than 2 meters tall) to the total area of the plot.
lign2_pct	LIGN2	Represents the ratio of the area covered by tall woody vegetation (over 2 meters tall) to the total area of the plot.
herb_pct	HERB	Represents the ratio of the area covered by herbaceous and mossy plants to the total area of the description plot.

EXPO is transformed from centesimal degrees to sexagesimal degrees (multiplying by 0.9) .
Additionally, we add new variables: coord_sys with the coordinate system employed (LAMBERT) and its CRS .

Variables regarding to herb (HERB) or woody vegetation (LIGN) are collected in the understory table.

The final variables provided are id_unique_code, province_code, class, subclass, plot, sp_name, sp_code, height and cover.

FIA (US): data processing

The Forest Inventory and Analysis (FIA) program, managed by the United States Forest Service, is a systematic effort to collect, analyze, and disseminate information about the nation’s forests. It provides comprehensive data on forest resources, including tree species, forest structure, composition, health, and dynamics.

FIA employs a cyclical approach to data collection, with periodic inventories conducted across different regions of the United States. These inventories typically occur on a 5 to 10-year cycle, although frequency may vary depending on the specific objectives and resource availability.

Data collection involves field measurements, remote sensing technologies, and statistical sampling methods to gather information on forested lands. Field crews collect data on tree species, size, density, growth, mortality, and other ecological attributes. Remote sensing data, such as aerial photography and satellite imagery, are also utilized to supplement ground-based observations.

Permanent plots in the FIA program are established using a systematic sampling approach, where sample locations are randomly selected within defined strata to ensure representative coverage of different forest types and ownership.

The Forest Inventory and Analysis (FIA) program, employs a phased approach to systematically monitor and assess forest resources across the nation.

• Phase 1 involves the initial setup of the sampling framework, including stratification of the landscape, random selection of sample plots, and establishment of permanent plot locations.

• Phase 2 consists of intensive field data collection on a subset of sample plots, focusing on detailed measurements of tree and vegetation characteristics. This phase provides comprehensive information on forest composition, structure, and health.

• Phase 3 encompasses specialized studies and additional data collection efforts beyond the standard inventory, addressing specific research questions or management objectives. This phase allows for targeted investigations into ecological processes, biodiversity, and environmental change.

In the Forest Inventory and Analysis (FIA) program, there are several core variables that are commonly collected across all sample plots. These variables provide fundamental information about forest resources and are essential for monitoring and assessing forest conditions. Some of the core variables include: tree attributes, stand attributes, forest health, vegetation composition and site characteristics.

Data acquisition

Data is obtained from FIA data mart where data can be downloaded by selecting the state(S) of interest or entirely. In any case it is organized by state in folder and it has archives with similar names except for the prefix.

https://apps.fs.usda.gov/fia/datamart/datamart.html

Here we use the following files:

“_TREE.csv”, “_PLOT.csv”, “_SURVEY.csv”, “_COND.csv”, “_SUBPLOT.csv”, “_VEG_SUBPLOT_SPP.csv”, “_SEEDLING.csv”, “_VEG_SUBPLOT.csv”, “_P2VEG_SUBPLOT_SPP.csv”

We also download this : “FIADB Reference Table CSV Archive” ( folder in zip format called FIADB_REFERENCE.zip ) and save REF_SPECIES.CSV and REF_PLANT_DICTIONARY.CSV.

Package has a version of these files obtained in march 2023 but you can provide an updated table.

Individual functions for FIA tables

Harmonized name	Original name FFI	Description
id_unique_code	US_STATE_COUNTY_PLOT	Code to identify plot
plot	PLOT	Plot ID
subplot	SUBP	Subplot number (up to 4)
year	INVYR	Year
county_code	COUNTYCD	Territorial division within states
state_code	STATECD	Territorial division

FIA TREE TABLES PROCESSING

This table provides information for each tree 1 inch in diameter (2.54cm) and larger that are found on a microplot, subplot, or core optional macroplot.

• Trees with diameters between 1.0- and 4.9-inches, termed saplings, are measured on the 6.8-foot radius microplot.

• Trees with diameters of 5.0-inches and larger are measured on the 24-foot radius subplots. Macroplot tree diameter thresholds are determined according to regional specifications (see regional field guides for more information).

Information is obtained from table “_TREE.CSV” .

Harmonized name	Original name FFI	Description
tree_id	TREE	Tree id
dbh (cm)	DIA (inches)	Diameter
height (m)	HT (feet)	Total height
density_factor (1/ha)	TPA_UNADJ (1/acres)	Expansion factor
sp_code	SPCD	Species code
status	STATUS	Dead or alive

Variables are transformed:

dbh from inches to cm by multiplying by 2.54

height from feet to m by multiplying by 0.3048

density_factor from acres to hectares by multiplying by 0.4046856422

Final variables provided are tree_id, status, dbh, height, sp_name, sp_code and density_factor.

FIA SHRUB TABLES PROCESSING

Shrub data is obtained from two different sources of information : “_VEG_SUBPLOT_SPP.csv” in case that there is available information in the file for the specific plot and year , as it is a protocol of P3 phase or “_P2VEG_SUBPLOT_SPP.csv” (P2 phase).

• Process for _VEG_SUBPLOT_SPP.csv (P3)

Harmonized name	Original name FFI	Description
subplot	SUBP	Subplot number
sp_code	VEG_SPCD	Species code
sp_name	-	Latin name from plant dictionary
cover	SP_CANOPY_COVER_TOTAL	Total cover of species in the subplot
height* = 91cm	SP_CANOPY_COVER_LAYER_1_2	Cover of layer between 0 and 6 feet 0 (0-1.8m)
height* = 340 cm	SP_CANOPY_COVER_LAYER_3	Cover of layer between 6 and 16 feet (1.8-4.9 m)
height* = 500 cm	SP_CANOPY_COVER_LAYER_4	Cover of layer above 16 feet (4.9 m)
growth_form	GROWTH_HABIT	Variable from Ref_plant dictionary: forbs, herb, shrub, graminoid …

Data is obtained at subplot level, but certain information ( default Height ) is aggregated at plot level.

We use maximum cover percentage of a species in different layers (SP_CANOPY_COVER_LAYER_X) to calculate the averaged height at plot level. For that, mid point of the interval is employed except for layer 4 where minimum is used.

Maximum cover	Default height
Layer 1-2	91 cm
Layer 3	340 cm
Layer 4	500 cm

All information is at subplot level.

Since P2VEG_SUBPLOT_SPP collects data from different growth forms (growth_form), we use the REF_PLANT_DICTIONARY to select data from shrubs ( code = “Shrub”).

Information about herbs is collected in a separated table (Herbs), growth_form= (“Forb/herb”, “Graminoids”) for these records.

Final variables provided are subplot, sp_name, height, cover, growth_form and sp_code.

• Process for P2VEG_SUBPLOT_SPP.csv (P2)

Data is obtained at subplot level, but certain information ( default height ) is aggregated at plot level.
Height is approximated by using the LAYER info.

Harmonized name	Original name FFI	Description
subplot	SUBP	Subplot number
sp_code	VEG_SPCD	Species code
cover	COVER_PCT	Total cover of species in a specific subplot plot and layer
height* (cm)	LAYER	VERTICAL LAYER LAYER 1 –> 0 to 2.0 feet LAYER 2 –> 2.1 to 6.0 feet LAYER 3 –> 6.1 to 16.0 LAYER 4 –>Greater than 16 feet
growth_form	GROWTH_HABIT_CD	Shrub (SH) , forbs (FB), graminoids (GR)

LAYER	Default height
LAYER 1	30 cm
LAYER 2	122 cm
LAYER 3	335 cm
LAYER 4	500 cm

All information is at subplot level.

If there is data in both files ( P2VEG_SUBPLOT_SPP and _VEG_SUBPLOT_SPP) , P3 data is prioritized and P2 data is added if gives additional information on certain species.

Final variables provided are subplot, sp_name, height, cover, growth_form, growth_form_code and sp_code.

FIA REGEN TABLES PROCESSING

Regeneration data (seedling) is obtained from : “_SEEDLING.csv” .

Seedling: live trees of a species found on a microplot that are less than 1 inch in diameter but at least 6 inches in length for conifer species or at least 12 inches in length for hardwood species.

Harmonized name	Original name FFI	Description
subplot	SUBP	Subplot number
sp_code	SPCD	Species code
-	TREECOUNT_CALC	Tree count used in calculations
density_factor	TPA_UNADJ (1/acres)	Expansion factor

TPA_UNADJ is transformed from acres to hectares by multiplying by 0.4046856422

New variables added:

• n is calculated by multiplying density_factor * TREECOUNT_CALC

• height = 15 cm (default value)

• dbh = 2.54 cm (default value)

Final variables provided are subplot, sp_name, treecount_calc, density_factor, n, height, dbh and sp_code.

FIA PLOT TABLES PROCESSING

Plot data (seedling) is obtained from : “PLOT.csv” , “SURVEY.csv”, and “COND.csv”

Data is obtained from both the last available year with the required information for the plot and the year of interest. If data for queried year is not available, we provide the last recorded value. Note that in most cases information on plot condition is collected once, when the plot is established.

Harmonization name	Original name	Description	Source table
rscd	RSCD	Region or station code	SURVEY
state_code	STATECD	State code	SURVEY
state_ab	STATEAB	State abbreviation	SURVEY
state_name	STATENM	State name	SURVEY
plot	PLOT		PLOT
county_code	COUNTYCD	State name	PLOT
year	INVYR	Inventory year	PLOT
coordx	LAT	The approximate latitude of the plot in decimal degrees. Actual plot coordinates cannot be released Therefore, this attribute is approximately +/- 1 mile and, for annual inventory data, most plots are within +/- ½ mile.	PLOT
coordy	LON	The approximate longitude of the plot in decimal degrees. Actual plot coordinates cannot be released. Therefore, this attribute is approximately +/- 1 mile and, for annual inventory data, most plots are within +/- ½ mile.	PLOT
elev	ELEV	The distance the plot is located above sea level.* For certain FIA work units (SURVEY.RSCD = 22, 23, 24, 33), the ELEV value is rounded to the nearest 10 feet. For other FIA work units (SURVEY.RSCD = 26, 27), the ELEV value is based on 200-foot groupings, and then a mid-point value is returned starting at 100 feet.	PLOT
p3panel	P3PANEL	Phase 3 panel number. A panel is a sample in which the same elements are measured on two or more occasions. FIA divides the plots in each State into 5 panels that can be used to independently sample the population. The value for P3PANEL ranges from 1 to 5 for those plots where Phase 3 data were collected. If the plot is not a Phase 3 plot, then this attribute is left blank (null).	PLOT
design_code	DESIGNCD	A code indicating the type of plot design used to collect the data. Refer to appendix G for a list of codes and descriptions	PLOT
p2veg_sampling_status_cd	P2VEG_SAMPLING_STATUS_CD	P2 vegetation sampling status code. A code indicating whether the plot is part of the P2 (Phase 2) vegetation sample included in the inventory. 0 Plot is not part of the P2 vegetation sample. 1 P2 vegetation data are sampled only on accessible forest land conditions. 2 P2 vegetation data are sampled on all accessible land conditions.	PLOT
p2veg_sampling_level_detail_cd	P2VEG_SAMPLING_LEVEL_DETAIL_CD	. A code indicating whether data were collected for vegetation structure growth habits only, or for individual species as well. Code Description 1 Data collected for vegetation structure only; total aerial canopy cover and canopy cover by layer for tally tree species (all sizes), non-tally tree species (all sizes),shrubs/subshrubs/woody vines, forbs, and graminoids. 2 Vegetation structure data (LOD = 1) plus understory species composition data collected including up to four most abundant species per GROWTH_HABIT_CD per subplot of: seedlings and saplings of any tree species (tally or non-tally) <5 inches d.b.h. (d.r.c. for woodland species), shrubs/subshrubs/woody vines, forbs, and graminoids. 3 Vegetation structure data, understory species composition data (LOD = 2), plus up to four most abundant tree species (tally or non-tally) <5 inches d.b.h. (d.r.c for woodland species) per GROWTH_HABIT_CD per subplot.	PLOT
aspect	ASPECT	Condition aspect. The aspect across the condition to the nearest 1 degree. North is recorded as 360. When slope is <5 percent, there is no aspect and this item is set to 0.	CONDID
slope	SLOPE	Condition percent slope. The predominant or average angle of the slope across the condition, to the nearest 1 percent	CONDID
-	CONDID		CONDID

We add id_unique_code and coord_sys : NAD 83 datum except for SURVEY.RSCD = 26 and SURVEY.STATECD = 60, 64, 66, 68, 69, or 70 that use WSG84 datum.

Elevation is transformed from feet to meters by multiplying by 0.3048.

Values of slope and aspect are divided into different conditions that plot may have (forest, non-forest) , as a proxy we use CONDID= 1 , that it is normally designated by the condition class at plot center (center of subplot 1)

FIA SUBPLOT TABLES PROCESSING

Subplot data is obtained from : “SUBPLOT.csv”

Harmonization name	Original name	Description
subplot_status	SUBP_STATUS_CD	Subplot/macroplot status code. A code indicating whether or not forest land was sampled on the subplot/macroplot. 1- Sampled - at least one accessible forest land condition present on subplot 2-Sampled - no accessible forest land condition present on subplot 3-Nonsampled - possibility of forest land.
macro_cond	MACRCOND	Macroplot center condition.
subplot_cond	SUBPCOND	Subplot center condition.
micro_cond	MICRCOND	Microplot center condition.
slope_subplot	SLOPE	Subplot percent slope. The predominant or average angle of the slope across the subplot, to the nearest 1 percent
aspect_subplot	ASPECT	Subplot aspect. The aspect across the subplot, to the nearest 1 degree. Aspect is measured by sighting along the direction used to determine slope. North is recorded as 360
psveg_subplot_status	P2VEG_SUBP_STATUS_CD	P2 vegetation subplot status code. A code indicating if the subplot was sampled for P2 vegetation. 1- Subplot sampled for P2 vegetation 2- Subplot not sampled for P2 vegetation

Final variables are subplot, slope_subplot, aspect_subplot, macro_cond, subplot_cond, micro_cond, subplot_status and psveg_subplot_status.

Appendix 1

#FIRST WE NEED A SPECIES LIST  , IT  IS OBTAINED FROM CSV METADONNES 
 library(GIFT)
#FINAL VECTOR OF SPECIES 
fr_species_cdref  

# growth form from GIFT
trait_meta <- GIFT::GIFT_traits_meta()
gf_trait_id <- trait_meta[which(trait_meta$Trait2 == "Growth_form_1"), ][["Lvl3"]]
growth_form <- GIFT::GIFT_traits(
  trait_IDs = c(gf_trait_id), agreement = 0.66,
  bias_ref = FALSE, bias_deriv = FALSE
)

growth_form_lignified_france <- growth_form |>
  dplyr::filter(work_species %in% fr_species_cdref) |>
  dplyr::select(work_species, work_author, trait_value_1.2.1) |>
  dplyr::distinct() |>
  dplyr::rename(
    GrowthForm = trait_value_1.2.1,
    AccSpeciesName = work_species
  )

head(forestables:::fr_species_cdref) 
#> [1] "Abies alba"            "Abies bornmuelleriana" "Abies cephalonica"    
#> [4] "Abies concolor"        "Abies grandis"         "Abies nordmanniana"

head(forestables:::growth_form_lignified_france)
#>       AccSpeciesName               work_author GrowthForm
#> 1         Abies alba                     Mill.       tree
#> 2      Abies grandis (Douglas ex D.Don) Lindl.       tree
#> 3 Abies nordmanniana            (Steven) Spach       tree
#> 4      Abies pinsapo                    Boiss.       tree
#> 5      Abies procera                    Rehder       tree
#> 6    Acacia dealbata                      Link       tree

References

ICONA (1990) Segundo inventario forestal nacional: Explicaciones y metodos. 1986-1995. Retrieved from: https://www.miteco.gob.es/content/dam/miteco/es/biodiversidad/servicios/banco-datos-naturaleza/explicacionesymetodos_tcm30-281101.pdf

IGN. (2014). Les données de l’inventaire forestier : état des lieux et évolution. Retrieved from: https://inventaire-forestier.ign.fr/IMG/pdf/LP_1.pdf

IGN. (2018). UN INVENTAIRE FORESTIER ANNUEL sur l’ensemble de la France métropolitaine. Retrieved from https://inventaire-forestier.ign.fr/IMG/pdf/180920_plaquette-methode_fr.pdf

IGN. (2022). MÉTHODOLOGIE Pour bien comprendre les résultats publiés 2017-2021. Retrieved from: https://inventaire-forestier.ign.fr/IMG/pdf/methodologie-2022.pdf

Forest Service USDA. (2023) The Forest Inventory and Analysis Database: Database Description and User Guide for Phase 2 (version 9.1) Retrieved from: https://www.fs.usda.gov/research/understory/forest-inventory-and-analysis-database-user-guide-phase-2

MINISTERIO DE TRANSICIÓN ECOLÓGICA Y RETO DEMOGRÁFICO. (n.d.). Cuarto Inventario Forestal Nacional (IFN4): Descripción de los códigos de la base de datos de campo. Retrieved from https://www.miteco.gob.es/content/dam/miteco/es/biodiversidad/temas/inventarios-nacionales/documentador_sig_tcm30-536622.pdf

TRAGSATEC (n.d.). 3er Inventario Forestal Nacional: Descripción de los códigos de la base de datos de campo. Retrieved from https://www.miteco.gob.es/content/dam/miteco/es/biodiversidad/servicios/banco-datos-naturaleza/documentador_bdcampo_ifn3_tcm30-282240.pdf