The Italian National Forest Inventory

Geographical and positioning aspects in relation to the different phases of the project Authors: Giacomo Colle Antonio Floris Gianfranco Scrinzi

Forest and Range Management Research Institute CRA - Agriculture Research Council - Italy 8th Annual FIA Symposium – Monterey 16-19th October 2006

Introduction to IFNI Second Italian National Forest Inventory “National Inventory of Forests and Forest Carbon Sinks” - INFC

Forest and Range Management Research Institute Agriculture Research Council

National Forest Service of Ministry of Agriculture and Forestry Policies

-

project study

-

scientific support

-

software applications analysis

-

data control and processing

-

project financing

-

project management

-

field surveys

ISAFA 2006 2

Introduction to IFNI Main aim of IFNI Providing updated and reliable information on the Italian forests to:

•

meet international reporting commitments (Kyoto Protocol, Convention on Biological diversity, Ministerial Conference for the Protection of Forests in Europe); •

direct Italian forest policies; •

collect information about forests at a regional level.

ISAFA 2006 3

Introduction to IFNI The Italian Scenario – 1st IFNI (1985) COUNTRY AREA : ~ 30,000,000 ha PERCENTAGE OF FOREST LAND : ~ 30% mixed coniferous

TREE SPECIES : hardwood

highforest

SILVICULTURE SYSTEM : coppices

public

OWNERSHIP :

private

ISAFA 2006 4

Introduction to IFNI Stratified Three - phase Sampling PHASE 1 - Land Cover classification 301,000 sample points randomly selected within each 1x1 Km square of the sampling grid ( in green: points recognized as forest) PHASE 2 - Forest Classification and qualitative attributes 30,000 sample points randomly selected within the stratum of forest and other wooded land. PHASE 3 – Quantitative attributes measurement ~ 7,000 sample points selected within the stratum of the second phase points. ISAFA 2006 5

Introduction to IFNI Stratified Three - phase Sampling

C

1 km

150 m

qua s g plin m a s

re

PHASE 1 : Photo interpretation

Year 2004

PHASE 2 : First ground survey

Year 2005

PHASE 3 : Second ground survey

Year 2006

r = 25 m Photo-interpretation and ground surveys have to be carried-out from the same position C

r = 13 m

ISAFA 2006 6

Contents of this presentation

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 7

Production of the national sampling grid

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 8

Production of the national sampling grid Aim : to determine the coordinates of the sample points in all the cartographic systems involved in IFNI

UTM / WGS84 : GPS positioning data of ground surveys

GAUSS-BOAGA / Roma40 : digital orthophotos land registers

UTM / ED50 : Italian official maps

ISAFA 2006 9

Production of the national sampling grid Aim 1: to determine the coordinates of the sampling grid nodes Aim 2: to randomly select inside each square the coordinates of sample points Statistical constraint: each square of the sampling grid must have the same area (1 km2) on the ground. Italy country crosses three UTM-WGS84 zones. The creation of a regular UTM square grid, starting from the central meridian of zone 32, will cause an excessive area distortion in the East regions of Italy.

The adopted solution is to create on the WGS84 ellipsoid a grid of “elliptic squares”, each one with the same area on the elliptic surface. As a result, a projected grid with “squares” of different shape is obtained.

ISAFA 2006 10

Production of the national sampling grid 1- The starting point of the procedure for the sampling grid creation is located in Monte Mario – Rome. This point determines the central meridian of the sampling grid. αI

β

α

1k

m

1 km

2- The first point is located 1 km East along the parallel passing through the central meridian. The angle α corresponding to this step, is function of the latitude of the starting point.

3- In the same way all the points located along the parallel are created. After that, a new point located 1 km North along the central meridian is generated. The angle β corresponding to this step, varies with the latitude.

4- Following the same procedure of step 2 (now with a new αI), all the points along the new parallel are created. 5- The procedure is repeated moving North and South along the central meridian. A grid as wide as Italy with same size elliptic squares, is created. ISAFA 2006 11

Production of the national sampling grid Sample points coordinates production 1 Inside each “elliptic square” of the sampling grid that overlaps Italy country, a couple of geographic WGS84 coordinates is randomly extracted. 2 The sample point coordinates are projected in each other IFNI cartographic system C planar coordinates in: UTM-WGS84

C geographic coordinates in:WGS84

C planar coordinates in: Roma40-Gauss Boaga C planar coordinates in: UTM-ED50

ISAFA 2006 12

Procedure for land navigation and GPS positioning

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 13

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 Aim 1 : to locate on the ground the sample point C Aim 2 : to collect data for the retrieval of point C during Phase 3

C

ISAFA 2006 14

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 GPS receiver and handheld configurations

Trimble Pro XR with Intermec handheld as data logger

Trimble Geo XT with external antenna

ISAFA 2006 15

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 GPS settings

Three possible settings of GPS filter values: Setting

Max PDOP

Min Elevation

Min SNR

Best accuracy

4

15°

6 AMU

Standard

8

12°

4 AMU

Best performance

12

12°

3 AMU

All GPS positionings have been generally performed with the best accuracy setting. Only in rare situations the best performance setting had to be used. Periodically during the campaign, the surveyors performed the planning to determine visibility for GPS satellites.

ISAFA 2006 16

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 Software in Phase 2

INFOR2 application has been developed for the Phase 2 of IFNI. The software runs on WinCE handhelds and provides “ad hoc” navigation and positioning functions. It also provides data storage functionalities for the qualitative attributes measurements of Phase 2.

ISAFA 2006 17

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 When the surveyor is far from C

Standard navigation procedure The distance and the bearing to C is obtained each time from a single GPS position. Printed digital orthophotos and cartographic maps help the surveyor to navigate towards the point C C

ISAFA 2006 18

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 When the distance to C is between 10 and 30 meters Sample point C is not a tangible object, it is a geographical position that has to be localized on the ground. F C

The procedure for the localization of C has to be as objective as possible.

The adopted method for the location of C, involves a second point, F, chosen subjectively. This method has been inspired by the ground sampling procedures adopted by the British Columbia Resource Inventory Committee (Canada) for the vegetation resources inventories. ISAFA 2006 19

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 When distance to C is between 10 and 30 meters Point F Point F is chosen around the point C, where the forest crowns allow the best transparency to the GPS signal.

F C

The coordinates of point F are obtained with an average GPS positioning of more than 170 positions. F location has to be chosen not too far from C, otherwise it will be inaccurate to locate in the forest the exact position of C using compass and tape measure.

ISAFA 2006 20

Procedure for land navigation and GPS positioning Navigation procedure in Phase 2 When distance to C is between 10 and 30 meters Standing in point F, point C is located on the ground calculating the distance d and the azimuth α from the GPS (average) coordinates of F to the theoretical coordinates of C

N α F

d

C

C

F

The distance d and the azimuth α from F to C will be used in Phase 3 to help the retrieval of the sample point

N α d ISAFA 2006 21

Procedure for land navigation and GPS positioning Positioning procedure in Phase 2 Get the coordinates of the centre of plot area in Phase 2

F

C

CGPS

Point CGPS After the (actual) C location has been individuated, its coordinates are measured using an average GPS positioning of more than 170 positions. These coordinates, called CGPS, differ from the theoretical coordinates of C, due to: • GPS positioning error on F and CGPS; • Error in the measurements of distance and azimuth

Theoretical coordinates of sample point C

ISAFA 2006 22

Procedure for land navigation and GPS positioning Temporarily marking the ground around C and F points in Phase 2 Aim : to help the retrieval of point CGPS in Phase 3 (performed about one year later) Constrain : to keep the invisibility of the places where the steel stakes of CGPS and F are situated FOBJ The object of F is usually a tree; it is marked with a painted cross and a little steel plaque fixed to the base of the tree facing F direction.

FOBJ F C C

C C

COBJ 2-3 little steel plaques are fixed to the base of an equal number of trees, each one facing C.

ISAFA 2006 23

Procedure for land navigation and GPS positioning Temporarily marking the ground around C and F points in Phase 2 Examples of marks in F and CGPS

F painted cross Painted cross on the tree chosen as F object

CGPS plaque Plaque fixed to a tree close to C

ISAFA 2006 24

Procedure for land navigation and GPS positioning Temporarily marking the ground around C and F points in Phase 2 Temporarily stakes are placed in F and in CGPS to retrieve the exact positions in Phase 3

F stake The steel stake in F holds a plaque with a hole placed in the direction of C. The hole distinguishes the F stake from the C one.

CGPS stake In CGPS is placed a temporarily steel stake with an aluminum plaque which is meant to ease the retrieval of the point during Phase 3, when a metal detector will be used. ISAFA 2006 25

Procedure for land navigation and GPS positioning Alternative procedure to the GPS navigation

P1

I

Open traverse If GPS signal is obstructed by the orography of the terrain or by a high canopy density, the IFNI procedure allows to use the traditional navigation method (open traverse).

P2 F

–

The open traverse originates at a starting point I, the last location along the path to C with good GPS signal conditions.

–

An average GPS positioning of more than 170 positions is performed in I.

–

The open traverse proceeds to the target point C with n stations P1, P2, …, Pn; where Pn is F.

C

The surveyors work with compass, laser rangefinder and clinometer. A special functionality of INFOR2 helps to calculate the coordinates of each traverse station and the distance and bearing to point C. ISAFA 2006 26

Procedure for land navigation and GPS positioning Alternative procedure to the GPS navigation INFOR2 include a set of functions to aid the surveyor in the open traverse procedure.

The software supports the calculation of the coordinates of each traverse station and of the distance and bearing to point C .

ISAFA 2006 27

Procedure for land navigation and GPS positioning Trimble SSF file recording Trible SSF File Coordinates and other info related to each GPS position are stored in a Trimble SSF file. The content of each SSF file has been extracted in a database for further elaborations.

ISAFA 2006 28

Analysis on GPS positioning data

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 29

Analysis on GPS positioning data Positioning analysis on SSF data extraction Valid SSF data of CGPS Valid SSF CGPS coordinates 74%

CGPS coordinates further than 20 m from C 0.24%

Approximately 74% of points have valid SSF CGPS coordinates for the analysis

SSF corrupted or missing 5% SSF with less than 170 GPS positions 21%

ISAFA 2006 30

Analysis on GPS positioning data Positioning results Percentage of CGPS points below distance D from C coordinates Distance of CGPS from C 120

Percentage

100 80

61

79

98 98 99

100 100 100

Totality

51 48

60 40

71 68

89 88 92

22 20

Geo XT Pro XR

28

20 0 1

2

3

5

10

20

Distance (m)

Pro XR performed better than Geo XT, even if it was used in unfavourable terrain and forest conditions.

Almost 90% of CGPS points is below 5 meters from C ISAFA 2006 31

Analysis on GPS positioning data Positioning results CGPS scatter plot around theoretical C coordinates

10

CGPS points are located uniformly around C.

Vertical position (m)

5

0 -10

-5

0

5

10

-5

-10

Horizontal position (m)

ISAFA 2006 32

Analysis on GPS positioning data Positioning results Percentage of CGPS coordinates obtained with a maximum Horizontal Dilution of Precision (HDOP) HDOP MAX in CGPS 120

Percentage

100

86

97

98

99

99

100

100

6.0

7.0

8.0

9.0

10.0

20.0

Out of range

72

80 52

60 40

93

26

20 0 2.0

3.0

4.0

5.0

HDOP MAX Value

Approximately half of the positioning have been performed with high accuracy values (HDOP ≤ 3) and almost the totality in standard conditions (HDOP ≤ 8) ISAFA 2006 33

Analysis on GPS positioning data Positioning results Comparison of HDOP in CGPS and in F Comparison of HDOP MAX in CGPS and in F 120

Percentage

100

81

80

91

96

98

99

99

100

100

100

100 80

62

60

60 40

120

35

40 20

20

0

0 2.0

3.0

4.0

5.0

6.0

7.0 HDOP MAX Value

8.0

9.0

10.0

Point CGPS

20.0

Out of range

Point F

The maximum HDOP value recorded during F positioning has been generally lower than in CGPS.

ISAFA 2006 34

Procedure for sample point retrieval in phase 3

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 35

Procedure for sample point retrieval in phase 3 Retrieval procedure of CGPS in Phase 3 Aim : to re-locate on the ground the stake placed in CGPS (in Phase 2)

In Phase 3 the sample point C is a tangible object (the stake placed in Phase 2). C

Therefore the procedure to localize C could both be successful or unsuccessful depending on the result of the CGPS stake retrieval.

ISAFA 2006 36

Procedure for sample point retrieval in phase 3 Retrieval procedure of CGPS in Phase 3 Software in Phase 3 NAV3 application has been developed for the Phase 3 of IFNI. The software runs on WinCE handhelds and provides “ad hoc” navigation and positioning functions. The surveyor could navigate towards any point of phase 2 using instant or average GPS positioning. NAV3 provides also a “moving map” navigation over digital orthophotos of Phase 1.

ISAFA 2006 37

Procedure for sample point retrieval in phase 3 Retrieval procedure of CGPS in Phase 3 Standard procedure 1- Navigate in the direction of point F using the GPS; CGPS F

2- Locate the object of F; 3- Using distance and azimuth from the object of F to F, locate the F stake; 4- Using distance and azimuth from F to CGPS and with the help of CGPS objects plaques, locate the C stake with the help of metal detector.

If F stake retrieval is successful, there is no need to do any average GPS positioning to re-locate the position where the stake of CGPS is placed. ISAFA 2006 38

Procedure for sample point retrieval in phase 3 Retrieval procedure of CGPS in Phase 3 Altenative procedures

CGPS

The recorded coordinates and the marks placed on the ground allow to find the C stake in different ways.

F

Example 1 • Navigate directly to CGPS using instant GPS navigation; • Using the marks facing to C, locate the stake of CGPS. Example 2 • Navigate directly to CGPS; • At the last step of the navigation perform an average GPS positioning to get accurate values of distance and bearing to CGPS; • Retrieve the CGPS stake with the help of metal detector. ISAFA 2006 39

Procedure for sample point retrieval in phase 3 Retrieval procedure of CGPS in Phase 3 If CGPS stake has not been retrieved If none of the procedures for the CGPS stake retrieval succeeds, the procedure used in Phase 2 to locate C from a point F has to be reexecuted in Phase 3.

F

CGPSI C

CGPS

The founded point CGPSI will need a new average GPS positioning of more than 170 positions. Neither a new object of F nor a stake in the new F are needed.

Theoretical coordinates of sample point C GPS coordinates of Phase 2

The positioning procedure of Phase 2 has to be repeated also if : • the SSF file of a sample point selected for Phase 3 is missed or corrupted; • its average coordinates comprise less than 170 GPS positions.

GPS coordinates of Phase 3

ISAFA 2006 40

Procedure for sample point retrieval in phase 3 Definitively marking the ground in CGPS

Phase 3 CGPS stake In sample points of Phase 3, a permanent stake is placed in CGPS (or CGPSI if the stake of Phase 2 was not retrieved). This stake will allow the retrieval of the sample point in the future. ISAFA 2006 41

GPS differential correction

Production of the national sampling grid Procedure for land navigation and GPS positioning Analysis on GPS positioning data Procedure for sample points retrieval in Phase 3 GPS differential correction

ISAFA 2006 42

GPS differential correction GPS real-time differential correction GPS real-time differential correction No GPS real-time differential correction has been performed during Phase 2 and Phase 3 surveys, this because neither WAAS-EGNOS nor other systems were still fully operative during the surveys. Moreover, the EGNOS system is difficult to use in mountainous land or with high canopy densities, due to the Inmarsat satellite position (azimuth 214°, elevation 33°). The real.time differential correction via mobile-phone was not possible, because the mobile-phone signal coverage is lacking in mountainous land if Italy.

ISAFA 2006 43

GPS differential correction GPS differential correction in post processing GPS differential correction A test on the European EUREF GPS Permanent Network has been performed on a subset of Trimble SSF files of Phase 2. Anyway surveyors of Phase 3 are easily retrieving CGPS points without using the differential corrected coordinates.

ISAFA 2006 44

THE END.

ISAFA 2006 45