Keywords

4.1 Introduction

Field surveys can be easily sketched by a sequence of basic, fundamental steps. In the great majority of cases, they consist of reaching the NFI sample point defined by its coordinates; assess the value of some variables; lay out sample plots (AdS) around or near it, within which to assess or measure other variables; and permanently mark the point, so as to be found in future field campaigns. The listed steps are also specific stages in the field campaign planning process, which are described in this chapter.

Navigation and positioning are determined through the use of global navigation satellite systems (GNSS); more specifically, special devices equipped with GNSS that are suitable to operate under dense tree cover and still maintain good positioning accuracy are employed. Sample point position is a fundamental variable assessed by NFIs, as it allows revisiting the sample plot and coupling ground data with remote sensing data for further analysis (Dalponte et al., 2011; Gobakken & Næsset, 2009; Kitahara et al., 2010). Assuring effective positioning for future retrievals of the sample point is done by marking it with either permanent buried markers or with other visible marks that help crews find the buried one, based on recorded information on their relative positions.

Land use and land cover classifications, the first carried out as preliminary to any other assessment, classification of vegetation and other qualitative characters as well as the measuring on the quantitative variables are the essence of NFIs. In this chapter, the methods adopted to survey the variables in INFC2015 are described. These are of general validity but adaptations of the protocol to special and rare circumstances are available from the field manual (Gasparini et al., 2016) or other cited references.

Recording the information collected in the field, either what is necessary for internal needs (e.g., those about the sample point markers) or necessary to produce statistics, is supported by computer tools for many reasons. These tools can make information more accessible and useful to surveyors (e.g., to validate or update information collected in the past field survey). They may help in more specific tasks (e.g., navigation and positioning). They may suggest logical order during the different steps of the survey, or ease data input by menu lists. They also allow for automatic checks, preventing entering inconsistent data or warnings in case the information entered is possible but dubious (possible inconsistency due to typing errors). Lastly, recordings allow easy electronic storage in the INFC central database simply by data transmission.

Data quality is another important aspect of field surveys. In high complex monitoring projects developed over years, good quality data cannot be based simply on a sequence of data collection, data check, data approval or rejection, as rejection implies severe consequences, with the work to be done again. It is important to support the crews while they are conducting the survey, not only to prevent mistakes and check the data in the central database to correct them, but also to recognize errors and prevent repetition of mistakes (Gasparini et al., 2009). Supporting the crews and carrying out quality checks on the database during the field campaign are activities to be taken into consideration during the planning of the field surveys, because they imply choosing appropriate communication tools and a proper database design.

4.2 Navigation, Positioning and Marking of Sample Points

Fig. 4.1
figure 1

GNSS navigation, positioning and data recording devices used during the field campaign of INFC2015 / La strumentazione per la navigazione, il posizionamento e la registrazione dei dati utilizzata durante i rilievi in campo INFC2015

Limiting the positional uncertainties of sample points in the different phases of the NFI, from photointerpretation to ground surveys, is a very important requirement for data quality. INFC adopts a three-phase sampling design. Phase 1 is carried out by photointerpretation and the other two phases use field surveys. Phase 2 uses about 30,000 sampling points, and phase 3 is a subsample of those. Photointerpretation is conducted in a sample area (photoplot 2500—FP2500) whose centre is the NFI sample point defined by its coordinates (cf. Chap. 2). The procedure to locate an inventory point on the ground with the highest possible accuracy in respect to its position in the interpreted orthophoto was developed in INFC2005 (Floris et al., 2011). This procedure was also adopted in INFC2015. INFC2005 also established marking procedures capable of ensuring retrieval of the sample points after a few months during phase 3. INFC2015 carried out a sole field campaign on sampling points that in 90% of cases had already been surveyed and marked in INFC2005. For this reason, it has been necessary to experiment and adopt a navigation and positioning strategy appropriate to maximise retrieval of existing marked points or establish new ones on the ground.

Navigation to the sample point (called C-point after its establishment on the ground) and its positioning were done using a multi-constellation GNSS receiver Trimble R1, operated by a datalogger Trimble Juno SB. Given the technological difficulties in using a real-time differential correction in many Italian forest areas, it was decided to use, whenever possible, a satellite-based augmentation system (SBAS) provided by the EGNOS service (European Commission, 2017; Gasparini et al., 2021). Approaching a position about 10–15 m distance from the sample point was carried out in instant position navigation (NPI) mode, while reaching specific coordinates (e.g., C-point established in INFC2005) from that position was carried out by a procedure called navigation from average position (NPM) (Colle et al., 2009). Figure 4.1 shows the devices used for navigation, positioning and data recording in the field.

The INFC sample points visited in the field during the previous inventory have a positioning uncertainty of 3–4 m, depending on the receiver used at the time (Colle et al., 2009). Those points were marked with two types of buried stakes: temporary stakes, for the points visited in INFC2005 phase 2 only, consisting of 30 cm long nail and a small aluminum head plate; and permanent stakes, with a steel tip with anchors and a larger aluminum head, for the points also visited in INFC2005 phase 3. Near the sample point, at approx. 10–15 m under the best local conditions for receiving the GNSS signal, an end of navigation point (F-point) was also marked through a buried temporary stake. Markings included nailing small aluminum plates to the base of three reference trees, whose monographic information recorded their species, DBH, distances and azimuth to C-point or F-point and photographs. In INFC2015, F-point and C-point stakes were searched by a metal detector, with the help of the monographic descriptions and the photographs taken in the previous inventory campaign. Examples of the marks are shown in Fig. 4.2. Figure 4.3 shows the regional and national rates of sample points for which at least one of the stakes (either in the F-point or in the C-point) was successfully found.

Fig. 4.2
figure 2

The buried stakes and the external markings used to mark the sample points in the field / I picchetti interrati e le marcature esterne usati per marcare i punti di campionamento in campo

Fig. 4.3
figure 3

Regional and national rates of sample points where at least one stake of INFC2005 (either point C or point F) was found in INFC2015 / Aliquote percentuali regionali e nazionale dei punti di campionamento nei quali almeno un picchetto di INFC2005 (punto C o punto F) sia stato ritrovato in INFC2015

In the INFC2015 field campaign, only permanent stakes were buried in C-points, in sample points in forest land use/land cover. To mark F-points, temporary stakes were used instead, considering the successful retrieval rate. At the end of the survey for each inventory point, a stationary GNSS positioning on F-point was conducted, calculating the coordinates as the mean of 50 single positions. Table 4.1 shows some positioning uncertainty parameters calculated from the GNSS raw files on more than 8000 points. As the C-point is reached measuring the azimuth and the distance from the F-point with a compass and a rangefinder, C-point and F-point positions undergo similar uncertainty. Finally, before leaving the sample plot, the monographic descriptions of the markers were updated and new photographs were taken.

Table 4.1 Positioning uncertainty parameters of sample points under different SBAS-EGNOS correction status / Parametri di incertezza del posizionamento dei punti di campionamento, in diverse condizioni di correzione SBAS-EGNOS

4.3 Variables, Their Classifications and Measurements

Field surveys allow definitive classification of land use and vegetation features relevant to verify correct inclusion of each sample point in the NFI domain and to classify the point in relation to a wide range of qualitative characteristics in order to measure quantitative variables. Table 4.2 lists the variables assessed or measured with the field survey and the reference sample unit (sample point or plot) used for each variable. Assessments and measurements are carried out with reference to the sample point or to one of the sample plots (AdS) that have it as the centre (AdS4, AdS13, AdS25) or close to it (AdS2) (cf. Chap. 2). This section describes the methods used for variable assessment or measurement. The classes adopted are described in the tables at the end of the chapters with the results, which show the statistics for the variables.

Table 4.2 Variables surveyed in the field in INFC2015, and their reference sample units / Variabili rilevate in campo in INFC2015 e rispettive unità di campionamento

4.3.1 Land Use, Land Cover, Pure/mixed Forests, Stand Origin

Crews first assess if the inventory sample point is in Total wooded area, by observing the land use and land cover (of trees and shrubs) status. The latter is quantified based on the orthophoto already used by the photointerpreter, available in the tablet. Surveyors judge potential crown cover, in case of temporarily unstocked areas, and accurately measure land features (e.g., roads and water streams width) when relevant for the adopted definitions in Chap. 2. A software application specifically developed in INFC2015, INFC_APP (cf. Sect. 4.4) guides assignment of a sample plot to the proper crown cover class. A systematic grid is superimposed to the orthophoto in the reference area (FP2500, cf. Chap. 2). Crews count the number of sample points falling on crowns and specify if these are shrubs or trees. Besides distinguishing Forest from Other wooded land, this assessment enables producing estimates on areas by total canopy cover, and by tree canopy cover. INFC_APP can also compute the predominance of conifer or broadleaf cover or neither (in mixed forests) to classify the stand as pure or mixed conifers/broadleaves.

Stand origin is assigned considering the intensity of human actions to promote or sustain regeneration; actions may be absent (natural stands), by silviculture (semi-natural stands) or intense (plantations).

4.3.2 Inventory Categories, Forest Types and Forest Subtypes

Inventory categories of Forest and Other wooded land are classified by observing features mainly related to composition (e.g., trees or shrubs), site potential for growth (e.g., tall trees or short trees), and stand origin (e.g., seminatural or planted). The inventory categories of Forest and Other wooded land are further classified into forest types and subtypes.

Forest type is classified based on the dominant species or group of species in terms of crown coverage. First, it is assessed if dominance is due to coniferous or broadleaved species, deciduous or evergreen; forest type must be consistent with that information, i.e., if crown cover prevalence is by conifer, one of the coniferous forest types is expected. In case of mixed stands, deciding which species or group of species is predominant may require a walk along with transects and recording of the species of the upper layer tree at established points (Fig. 4.4). Such procedures and computations are assisted by INFC_APP. Forest subtype is assigned based on the species composition and stand ecological characteristics.

Fig. 4.4
figure 4

Walking transects across AdS25 to assess tree species crown cover, by standing in the positions indicated by the circles. Red numbers show the order to follow to obtain at least 25 observations / Disposizione dei punti di osservazione della copertura delle specie entro AdS25; i numeri in rosso indicano l’ordine dei transect fino al raggiungimento di un totale di 25 classificazioni

4.3.3 Legal Status

Ownership was assessed based on the position of the inventory sample point, accessing cadastral information available to crews as a GIS layer, or by interviewing local forestry or other administrative personnel.

Limitations on the use of forest resources or the inclusion in protected areas (type and protection level) were assessed on documents or maps or given by local personnel.

Presence of forest planning was assessed based on documents or local personnel knowledge; planning may be present at various levels (e.g., regional guideline plans or property-based operative management plans).

4.3.4 Accessibility, Presence of Roads and Site Features

A sample point is accessible if the crew can reach it and can carry out the required assessment and measurements in the relative AdS. Physical conditions, prohibited access under laws or from owner’s decisions and risk of severe injury for the surveyors may result in inaccessibility. Flexibility is allowed during the survey in challenging cases. For example, if it is too risky, DBH of some trees might be visually estimated or the measurements limited to AdS4. Any special adaptation is recorded to allow consistent data processing.

Simple and fast access to a sample point also depends on the presence of roads. This is first assessed on maps, orthophotos, etc. and then checked and verified in the field. Crews record distances and differences of elevation between the NFI sample point and the closest road or forest track.

The inventory sample point altitude is measured by the GNSS device, while recording the coordinates (cf. Sect. 4.2).

Aspect is measured using a compass, by reading the magnetic azimuth; the measurer stands at the sample point aiming the compass downwards. The slope angle is measured by a clinometer. The measurer stands at the sample point and aims towards two topographic poles, placed 25 m distance uphill and downhill along the maximum slopes; the recorded value is the average of the two measures.

Land position is observed by referring to both the AdS25 (local land position) and to a larger area, from a few hectares to a few tens of hectares around the plot (extended land position). Roughness, assessed in the AdS25, is the micro-morphology of the terrain, determined by the presence of obstacles such as boulders, rocks, ditches and sinkholes that could condition any logging operations (felling, concentration, removal) and, in general, the availability of the area around the sample point.

Any presence of terrain instability under way, repetitive or occasional, is surveyed in the AdS25 or the surrounding area. The instability must affect an area of at least 100 m2 to be recorded.

4.3.5 Silviculture, Stand Characters, Production

Applied silviculture considers type and intensity of practices. These are strictly related to the products that are obtained, so the classification system has specific classes for the practices aiming at obtaining secondary, non-wood products and social services.

When the management is strongly and clearly oriented to mainly obtain one specific good or service from the forest, this is classified under the primary designated management objective condition, and the specific good or service is noticed.

Availability for wood supply is information provided based on documents (e.g., restrictions by environmental protection laws) or assessing in the field eventual economic convenience of utilisation or presence of severe constraints. Convenience of utilisation might come by considering, for example, the value of the exploitable wood in relation to the difficulties and costs in accessing remote forests.

The silvicultural system and the utilisation mode are recorded when observing stand structure and characters, while the logging mode also considers the exploitable products (e.g., timber or firewood), local tradition (e.g., cables are mainly used in the Alps and less on the Apennines) and road presence.

Development stage information is based on visual assessment as well as the age class, which is recorded only in even-aged stands; however, the information is confirmed after the tree coring described in Sect. 4.3.8, and once tree rings have been counted.

4.3.6 Forest Health

The survey aims to provide a general overview on forest health conditions and records the presence of diseases and damages when they affect at least 30% of AdS25. When this threshold is reached, the cause of the disease or damage is assessed as well as its severity on the trees that are sick or damaged. In case the disease or damage implies defoliation, its intensity is recorded, and an indication is given about which part of the crown is primarily affected.

4.3.7 Tally Trees

Tally trees are woody plants, either alive or dead, with DBH ≥ 4.5 cm inside the AdS4 or ≥ 9.5 cm in the AdS13. A tree is in or out of a sample plot depending on the distance between the plot centre and the vertical axis passing through the centre of the section at 1.30 m aboveground level. DBHs are measured using a calliper. On slope terrain (>10°) the measurer stands uphill from the tree, otherwise the graduated beam is along the plot radius. When the first measured DBH exceeds 9.5 cm, its cross-section diameter is also measured. For each callipered tree, the species, vitality and integrity (Table 4.3) and type of tree (Table 4.4) are also recorded. In case of broken trees, the height is measured. For dead trees, the decay class is recorded, based on visual assessment (Table 4.5).

Table 4.3 Tree vitality-integrity / Classi per la vitalità-integrità degli alberi
Table 4.4 Type of tree / Dendrotipo
Table 4.5 Decay classes adopted for deadwood assessment (tally trees, stumps, deadwood lying on the ground) / Classi di decadimento per il legno morto grosso (alberi morti, ceppaie, legno morto grosso a terra)

4.3.8 Sample Trees

Ten trees among the tally trees are selected for additional measurements. These include the five closest trees to the plot centre, the three remaining largest and two ‘rare’ trees among the remaining. A rare tree may be a tree of species or size class (in monospecific stands) not included within the previous eight. Whenever possible, sample trees must be free from visible faults. For each tree, the total tree height and crown base height are measured, and the tree is classified as dominant, intermediate or dominated tree. Tree height is measured using a Haglof Vertex hypsometer.

In order to estimate the annual diameter increment, the sample trees are cored with a Pressler increment borer, at 1.30 m aboveground level, along with the plot radius direction. From each sampled tree, one core is extracted and the length of the five outermost rings (excluding the one from the current year) is measured with a ruler. Each core is labelled and sent to the dendrochronology laboratory at CREA Research Centre for Forestry and Wood in Trento for a measurement check.

4.3.9 Forest Understorey

Small trees and shrubs are distinguished based on the INFC species list. Distinction is not relevant for measurements, since what must be measured only relies on size thresholds. Small trees and shrubs are woody entities with a diameter of less than the calipering threshold of 4.5 cm but higher than 50 cm. They are assigned to one of the classes in Table 4.6. They are measured in two AdS2, 10 m distance from the sampling point, East and West positioned (cf. Chap. 2). The survey consists of counting the plants separately by species and size class. For each species surveyed, the prevalent origin is recorded (artificial, agamic, by seed), eventual damage, if affecting at least 30% of the plants, and the cause, if recognisable (by animals, e.g., by pasture or wildlife, or by weather).

Table 4.6 Size classes for small trees and shrubs / Classi dimensionali per la rinnovazione e gli arbusti

4.3.10 Deadwood Lying on the Ground and Stumps

Deadwood lying on the ground (complete trees, stems, branches, etc.) must have a diameter and length of at least 9.5 cm and is measured within the AdS13. Each woody piece (called an element) is ideally divided into regular fragments of length not more than two metres. For each fragment, two cross sectional diameters at both terminal sections are measured, as well as its length. Each fragment is classified as either from coniferous or broadleaved species and the decay condition is given according to the classes in Table 4.5.

Stumps are the remains of cut trees or naturally broken trees not reaching a height of 1.30 m and with diameter at least 10 cm and are measured in the AdS13. The diameter at the cutting height (two orthogonal measurements) and the height aboveground level (two measurements, the minimum and maximum height) are measured. The species is also recorded, if recognisable, as well as the decay class (Table 4.5). Lastly, information is given on the cutting age, specifically if the cut has occurred before or after the twelve months preceding the survey.

4.4 Data Collection, Database and Field Software (INFC_APP)

The relational database designed to store in central server data collected during the field surveys was developed in an Oracle 10 g environment and consisted of 27 tables (more than 280 fields) linked by the sample point identifier as the primary key and, in the quantitative variables tables, by the item (tree, stump, etc.) identifier as the secondary key. Accessing the central server is possible through authentication and protection protocols, which vary in relation to four profiles associated with corresponding specific roles of the user. These include CUFA, which is nationally responsible; researchers/analysts of CREA Research Centre for Forestry and Wood, regional coordinators, and the head of the field crew. The database also stores and makes available the data recorded during the photointerpretation, which must be validated or updated during the field survey.

The device used for data storage in the field is a tablet with OS Android 6.0, for which the application INFC_APP was developed (Gasparini et al., 2020). Most of the preloaded data is contained in this app, except the waypoints for navigation and raw positioning files, which were recorded with Trimble Terrasync software on the Trimble Juno SB datalogger. On the client side, the database was implemented through the Sqlite library. The information flow between the tablet device (client) and the central server can occur directly from the field, using the installed G4 card, via web service Simple Object Access Protocol (SOAP) (Oracle, 2021). With this protocol, it is possible to transfer not only alphanumeric data, but also binary data as the .ssf files containing the raw measurements and the qualitative parameters of the GNSS point positioning. The client device sends a request and waits for a response from the server. Four web services were implemented for searching, downloading, uploading, and updating the status of the points, fundamental for the distinction between the points to be surveyed, those in progress or those that have been concluded. The web service receives the tablet’s IMEI code and returns the list of points assigned to the entitled NFI crew. The crew team can then choose the sample point to survey from that list and download all the preloaded information. If the procedure ends correctly, the client device automatically invokes the service, and this changes the status of the point, moving it from ‘assigned’ to ‘in progress’, thus excluding it from the list of points that are still possible to download on devices. At the end of the survey, the client invokes the service for uploading data, sends all the collected data to the central database, making them visible to other user profiles authorised to access and consult. It is possible to do partial uploads, as temporary backups of a survey not yet concluded, or a final upload to transfer all data at the end of the surveys on a sample point. The final upload can even be postponed to when data collection has been accomplished, such as after measurement of incremental cores, which is usually performed in the office.

Data security during the campaign is ensured by the establishment of a list of devices authorised to install the app and access the database according to the specific IMEI code of each tablet. In regard to INFC_APP development, one of the basic requirements during the design phase was to create a user-friendly GUI application. The user interface guides the surveyor throughout the logical and chronological phases of the survey, from navigation and positioning procedures to the collection of the qualitative and quantitative variables. It is possible to move through the various sections of the application, with the only constraint being to save the data entered in any specific section by explicitly confirming the saving. From the home screen it is possible to consult the documentation (survey protocol, electronic devices manuals, etc.) remaining in the application. The GUI sections are of two different types: the first is represented by data input modules (Fig. 4.5), and the second is the result of data queries, which resume the previously input data, both within each section and in the home screen, thus facilitating the monitoring of the survey progress (Fig. 4.6).

Fig. 4.5
figure 5

Example of INFC_APP input module, regarding attributes referring to tree callipering / Esempio di modulo di input di INFC_APP, riguardante attributi relativi al cavallettamento degli alberi

Fig. 4.6
figure 6

Example of INFC_APP output module, regarding the progress status of tree callipering / Esempio di modulo di output di INFC_APP, riguardante lo stato di avanzamento del cavallettamento

For all qualitative variables that are categorical and assessed using the appropriate class value, look-up tables have been adopted and the chosen class is selected from a closed list of possible values. For quantitative attributes, threshold values have been established, and warning messages have been shown in case of unlikely values. Combinations of masks and sub-masks have also been created for the input of progressively more detailed data (Fig. 4.7). Several real-time automatic cross-checks in different fields have been designed to prevent errors in terms of plausibility, congruity, completeness, or input errors (Fig. 4.8).

Fig. 4.7
figure 7

Example of INFC_APP module, with mask and sub-mask, regarding the input of progressively more detailed data status / Esempio di modulo di INFC_APP con maschera e sotto-maschera per l’inserimento di dati a dettaglio progressivamente maggiore

Fig. 4.8
figure 8

Two examples of completeness and plausibility real-time checks in INFC_APP / Due esempi di controlli di completezza e plausibilità in tempo reale in INFC_APP

4.5 Start-Up, Remote Assistance and Quality Check

4.5.1 Start-Up and Remote Assistance to Field Crews

As soon as the crews began their job, they were joined by the CREA Research Centre for Forestry and Wood team to be led under expert assistance. This assured appropriate reminder of the protocol procedure as learned in the training course and also provided clarification on local specific cases. However, clarification on local specific cases is not a secondary concern in Italy, given the highly diverse vegetation conditions.

During the field campaign, a helpdesk was active daily at CREA Research Centre for Forestry and Wood in Trento, reachable either on the phone or by email. Crew support consisted of answering questions related to any aspect of the survey protocol in an effort to help resolve doubts and uncertainties and so avoid subjective or incorrect interpretations. In fact, questions were often asked about special cases because crews were not in ordinary conditions, e.g., they were at risk of injury. All questions were entered in a database in such a way to give future consistent answers in similar cases, but especially to derive statistics on the most frequent topics of questions. Based on these statistics, the helpdesk could warn crews (all crews or only those potentially interested) to prevent reiteration or possible mistakes. The helpdesk database also allowed to verify full functioning of the automatic controls of INFC_APP, especially the first releases (cf. Sect. 4.4), but above all, they could suggest further improvements.

4.5.2 Data Quality Check and Final Field Work Checks for Approval

Regular checks on the central database are fundamental to guaranteeing high data quality. They also permit monitoring the progress of the field work, revealing possible problems in carrying out the survey which cause delays. In this case, they allowed verification of the crews’ accomplishments of what was requested about data and materials. Crews had to post the woody cores to the dendrochronology laboratory at CREA Research Centre for Forestry and Wood in Trento, as well as safety backup copies of the photographs taken. Arrival of these two additional materials was expected at stated intervals after sample plots data were sent to the central database.

Checks in the database during the fieldwork aimed at assessing completeness, plausibility, and consistency of information. Completeness indicates that all data expected have been recorded. This is possibly the simplest check to be implemented through automatic checks in the crews’ software; nevertheless, control is needed to verify the proper functioning of the software. In our case, controls revealed malfunctioning of early releases of INFC_APP, which caused a limited number of data loss. Checks about materials sent to CREA Research Centre for Forestry and Wood in Trento were special cases still related to completeness checks.

Plausibility means that information is reliable in absolute terms. An example of qualitative information that is not credible is the declared presence of a species in an area where it cannot be found (e.g., cork oak on the Alps). Another example involving quantitative values is a recording size that cannot be reached (e.g., exaggerated tree heights). The two examples should highlight the limits of automatic checks and importance of judgment based on data control, because there are places where a species is not expected but would still be possible to be found as well as a tree height that is possible may not be in a specific stand.

Consistency means that two pieces of information on related variables make sense; in other words, the two pieces of information are conditioned with each other. Nevertheless, the two values provided are reliable. Plausibility checks and consistency checks were carried out through cross-check data within the INFC2015 database, but they could also rely on crossing information from the INFC2005 survey.

Final field work checks for approval after the end of the field campaign was carried out by controllers. They fully surveyed some plots per region, in the presence of the crews and the data gathered were compared with those stored in the central database, previously recorded by the crews. The comparison was aimed to compute the level of reproducibility. For each variable measured, a maximum proportion of disagreement was stated, and a score was assigned, based on a level of agreement. The overall score was based on the scores obtained in each variable assessed. Crew knew from the very beginning that their job would be evaluated. For this reason, although limited in numbers, these checks were effective in maintaining a high level of quality throughout the entire field campaign.