the natural areas of rome province detected by airborne ... · nies with large land holdings...

187

ANNALS OF GEOPHYSICS, VOL. 49, N. 1, February 2006

Key words airborne hyperspectral remote sensing– environmental monitoring – maximum likelihoodvegetation mapping

1. Introduction

The importance of mapping, quantifying andmonitoring changes in the physical characteris-tics of forest cover has been widely recognizedas a key element in the study of global change(Nemani and Running, 1996). Land cover, i.e.the composition and the characteristics of land

surface elements, is a piece of environmental in-formation and represents an effective resourcefor management and policy purposes for a widerange of human activities. The location and ratesof forest structural change and the degree towhich landscapes respond to human distur-bances require extensive investigations (Lam-bin, 1998; Borak et al., 2000). Among the largevariety of forested ecosystems the Mediter-ranean comprises a vegetation form and distri-bution with worldwide importance being of spe-cial interest to global change research in termsof biodiversity, desertification, water resourcesand urbanization (Hill et al., 1995; Moreno andOechel, 1995).

The new generation high resolution opticalsensors allow new applications in Earth obser-vation for land cover mapping by deliveringnew methodologies for generating land cover

The natural areas of Rome Provincedetected by airborne remotely sensed data

Rosa Maria Cavalli (1), Lorenzo Fusilli (1), Anna Guidi (2), Simone Pascucci (1), Stefano Pignatti (1)(3),Ludovico Vannicelli Casoni (4) and Maria Vinci (2)

(1) Laboratorio Aereo Ricerche Ambientali (LARA), IIA-CNR, Tor Vergata (RM), Italy(2) Provincia di Roma - Dipartimento II, Roma, Italy

(3) Istituto Metodologie Analisi Ambientale (IMAA), CNR, Tito Scalo (PZ), Italy(4) Provincia di Roma - Dipartimento I, Roma, Italy

AbstractRome Province with its 4 million inhabitants is one of the Italians areas with the largest urban expansion, main-ly concentrated around the capital city. The uncontrolled urbanization of the past has heavily marked the land-scape, especially Rome countryside and coastline. However many zones have exceeded the anthropic pressure-without serious consequence since the sensitivity towards environmental protection has grown in recent years.Rome Province Administration has devoted special attention to the improvement and protection of its naturalisticheritage by means of a series of administrative actions, cultural initiatives and projects for environmental educa-tion. In this perspective a three-year agreement was concluded with CNR LARA focused on the study of naturalvegetation by means of MIVIS (Multispectral Infrared Visible Imaging Spectrometer) remotely sensed data. Thisstudy distinguished and mapped the most important natural forests, shrub and herbaceous formations, assessedthe health conditions of the arboreal vegetation, identified the areas with little water supply, and measured someenvironmental parameters, like temperature and surface humidity. The results achieved highlight the large botan-ical and naturalistic assortment and the complexity of the study-area.

Mailing address: Dr. Lorenzo Fusilli, Laboratorio Ae-reo Ricerche Ambientali (LARA), IIA-CNR, Via del Fossodel Cavaliere 100, 00133 Tor Vergata (RM), Italy; e-mail:[email protected]

188

Rosa Maria Cavalli et al.

information products with high accuracy. Forthe remote sensing community vegetation map-ping is a considerable objective for the studyand monitoring of ecosystems. The rapid andcost-effective application of remote sensing tomap vegetation is one of the important motiva-tions for its utilization in land use planning toreplace more time consuming and expensivefield surveys. In this context, an increasingnumber of agencies as well as private compa-nies with large land holdings currently use veg-etation maps derived from satellite and airbornedata (Congalton et al., 1993).

Rome Province Administration has long de-voted its efforts to enhancing the value and pro-tecting its historical naturalistic heritage andhence it has implemented some specific LandManagement Plans for its Natural Reserves(Nature 2000 Project of the European Commu-nity). The purpose of these Management Plansis to represent a reference framework fromwhich to undertake land management policy ac-tions and stimulate bio-compatible activities inthe respect and the safeguard of the environ-ment. They also aim to organize all geographic,urban and environmental data in a homoge-neous context, to carry out a careful analysisand intervention planning. The necessity hasthus arisen to make a detailed «Map of theMain Vegetation Classes» (scale 1:10000) ofthe Natural Reserves recently established, andof the north-eastern area of Rome Province ter-ritory (scale 1:25000) where a detailed homo-geneous cartography of vegetation has neverbeen compiled.

The considerable extension of the area to in-vestigate, the difficulty in approaching inacces-sible places, together with the biodiversity ofvegetation, led to the use of airborne remotesensing as a means of investigation. The cartog-raphy deriving from the Corine Land Cover at1:100000 scale (1994) obtained from Landsatsatellite images (ground pixel resolution of 30m) was not able to meet the required carto-graphic and thematic details. Rome ProvinceAdministration therefore started a triennialagreement with CNR-IIA/LARA to study themain vegetation covers by means of MIVIS(Multispectral Infrared Visible Imaging Spec-trometer) airborne sensor.

2. Study area

The area investigated is of particular interestfor both its botanic-naturalistic variety still in-tact and its historical-archaeological heritage.Local authorities, in particular Rome ProvinceAdministration, are properly improving andsafeguarding this area by means of administra-tive provisions, cultural initiatives and projectsfor environmental education.

The area (fig. 1), located in the E-NE por-tion of Rome Province (Latium), extends fromthe Tiber Valley and the lightly rolling undula-tions of the Roman countryside, towards thesharp carbonate massif near Tivoli, and east tothe mountainous ridge of the Latium Apen-nines. This territory has already been safeguard-ed to a great extent by establishing several pro-tected areas, some important like the NaturalReserves of Nomentum, Macchia di Gattaceca,Mt. Catillo, Mt. Soratte, and the Regional Parksof Mts. Lucretili and Mts. Simbruini. The highenvironmental value of this area is also provedby the presence of numerous Sites of EuropeanCommunity Interest (SIC) and Special Protect-ed Areas (ZPS) instituted according to the92/43/CEE «Habitat» and 79/409/CEE «Birds»provisions.

From the point of view of flora, the areacomprises several wood formations which areone of the prevailing land covers. Besides thearboreal component, the herbaceous and shrub-by formations, presenting a wide ecological andfloristic differentiation, are also well represent-ed. According to the Latium Phytoclimate Map(Blasi, 1994), the above plant formations fallinto the following phytoclimatic units:

– Temperate bioclimatic region: unit 2,summits of lower carbonate massif with lowermontane thermotype and upper humid om-brotype; units 3 and 4, Apennines intramontanevalleys with submontane thermotype and fromupper humid to lower hyperhumid ombrotype;unit 6, Sabina piedmont with hilly thermotypeand from subhumid to lower humid ombrotype.

– Temperate bioclimatic transition region:unit 7, Middle Tiber valley between Orte andMonterotondo with thermotype from hilly toMesomediterranean and inferior humid om-brotype.

189

The natural areas of Rome Province detected by airborne remotely sensed data

The whole area shows formations belongingto the Mediterranean environment, like ever-green forests with a prevailing presence ofQuercus ilex, and the typical Mediterraneanmaquis. This Mediterranean maquis which insome spots (mainly in the area of Mt. Catillo) ischaracterized by the peculiar chorologic easterncomponent with an abundant presence of Styraxofficinalis (protected species by the RegionalLaw 61/74), Cercis siliquastrum, Paliurusspina-christi, Carpinus orientalis (Montelucci,1972, 1984).

In the submontane climatic type (between200 and 600 m a.s.l.) deciduous species mixedwith Quercus pubescens, Quercus cerris andOstrya carpinifolia are present, at mountain lev-el (between 600 and 1600 m a.s.l.) Fagus sylvat-ica is present with a fragmentary distributionwhile in mountainous areas it is thickly spread(Montelucci, 1978). The articulated morphologyand diffuse surface karstification bring about thepresence of humid depressions covered by mes-ophyll prairies with different floristic types.

Arid prairies are diffuse as well, being the con-sequence of the prolonged exploitation ofmountain pastures; they represent the mostfavourable habitat for the growth of those vege-tation species typical to arid environments oncarbonatic lithologies. Rocky environments areof particular interest since they also host numer-ous endemisms and botanic rarities.

3. Data and processing methods

MIVIS data campaign, conducted in June1998, covered an area of about 168225 ha, cor-responding to about 31% of the Province’swhole surface. It included 28 runs recorded withrelative altitude of about 1700 m, NNW/SSE-oriented flight lines, parallel to one another, withapproximately 25-30% overlapped laterally (fig.2). The pixel ground resolution, due to the rela-tive flight altitude, is about 3.4 m.

The MIVIS is composed of 4 spectrometerswhich measure the electromagnetic Earth radia-

Fig. 1. Location of the study area on Rome Province with MIVIS swaths overlaid.

190


tion with 102 spectral bands in the visible, nearinfrared, middle infrared and thermal infrared re-gions, with 2 mrad of IFoV (see table I).

The dataset used was the first 28 bands (Iand II spectrometer) covering the spectral range0.43-1.55 nm because this spectral region cor-responds to the maximum reflectance of vege-tation and encompasses the major spectral fea-tures related to the biophysical characteristicsof the plants. The bands of the III and IV spec-trometers were not employed because they did

not produce any meaningful improvements inthe classification process and they increasedprocessing time.

MIVIS data have been radiometrically cali-brated using references inside the sensor, andthe Radiance Factor was measured on opticalbench; this operation converted the sensor’selectromagnetic radiation, expressed in DigitalNumbers (DN), into radiance [nW/(cm2 nm sr)].MIVIS images were then corrected for the pathradiance atmospheric component by means of

Table I. MIVIS spectral characteristics.

Spectrometer Spectral region Bands number Spectral range (nm) Band width (nm)

I Visible 20 0.43-0.83 0.02II Near infrared 8 1.15-1.55 0.05III Medium infrared 64 2.0-2.5 0.009IV Thermal infrared 10 8.2-12.7 0.34-0.54

Fig. 2. Mosaic of MIVIS flight lines.

191


simulations performed with the MODTRANcode (Berk et al., 1989), since such effect on ascene produces classification errors.

The geometric correction was carried outthrough a software developed in IDL languageby LARA (Avanzi et al., 1997, 2006) which en-abled us to correct the airborne attitude plat-form during flight for panoramic distortions,(due to yaw, pitch, roll, drift, altitude and speedvariations) and the morphological variations ofthe Earth surface. These corrections (fig. 3)were performed taking into consideration both

the system acquisition geometry and the air-borne navigation data. The topographic varia-tions of the countryside territory were insteadcorrected by means of the Digital ElevationModel (DEM) derived from satellite interfer-ometry techniques with a resolution of 30 m(courtesy of ESA-ESRIN). The residual errorsdue to the not perfect reconstruction of the air-plane’s trajectory and ground elevation wereminimized using the Ground Control Points(GCP) extracted from the Regional TechnicalMap (scale 1:10000). The image warping was

Fig. 3. Geometric correction: (left) uncorrected image; (right) geometrically corrected image.

192


accomplished in one single phase by means ofa «Nearest Neighbour» re-sampling.

This pre-processing chain methodology wasdeveloped and then tested on the four NaturalReserves, as their vegetation complexity consti-tutes a suitable test bench to experiment andimprove the different techniques adopted.

In view of the botanic and naturalistic im-portance of these Natural Reserves, the Map ofPrevailing Vegetation Classes was produced at1:10000 scale (Cavalli et al., 2001), while thecorresponding map referring to the entire air-borne remotely sensed area is at 1:25000 scale.In addition to the Map of the Main VegetationClasses, other specific thematic maps have beenproduced concerning some environmental andphytosanitary indicators like soil humidity, dryvegetation, NDVI (Normalized Difference Veg-etation Index) and NDWI (Normalized Differ-ence Water Index; Gao, 1996).

4. Implementation of the Map of the MainVegetation Classes

To manage the great quantity of available da-ta as best as possible, a relatively simple method-ology has been defined, which could be repro-ducible and applicable to all remotely sensedruns. In particular to reduce data processing time

and minimize the drawbacks occurring duringimage mosaiking, the area was divided into 3sectors, each one with its own morphologic andenvironmental characteristics. The first sector isconstituted by the mosaic of runs 1 to 7, charac-terized by mainly plain and hilly areas. The sec-ond one is formed by the mosaic of the runs 8 to18, mostly characterized by sub-mountain andmountain ranges. The third one comprises themosaic of runs 19 to 28, especially characterizedby mountain and high relief surfaces. These are-al divisions also minimized the problem of thenon uniform distribution of ground truths in allthe scenes for each spectral class.

To facilitate the image supervised classifica-tion, a first discrimination of the main homoge-neous land cover classes was defined by photo-interpreting MIVIS images (see table II) to cre-ate uniform spectral classes.

Informative layers were then produced bydigitizing on screen their limits (so obtainingvector files), covering the following classes:arable lands, permanent crops, urban areas,quarries, waters, natural environments. Vectorlayers so obtained were transformed into masksand used to create thematic subsets over whichthe usual techniques of image processing can beapplied for the supervised extraction of spectralclasses. In particular to perform the «Map ofPrevailing Vegetation Classes», a mask of natu-

Table II. False Color Composites (FCC) and Black and White (BW) images used for photo-interpretation. A-com-binations of bands corresponding respectively, for vegetation, to the two bands with higher spectral absorption (redand blue colours) and the one with major reflectivity (green colour): it better approaches the natural colour repre-sentation; B-combinations of channels corresponding respectively to the maximum reflectance plateau in the Near-Infrared, the water absorption peak and chlorophyll peak; hence, this FCC highlights vegetation covers, discrimi-nating between agricultural lands and spontaneous vegetation; C- and D-BW images which allow to emphasize lesseasily detectable covers such as bare soils, vegetation sparse or in relative dryness conditions.

FCC Red Green Blue BW images

A Ch. 13 Ch. 7 Ch. 1 (0.6 nm<m<0.69 nm) (0.53 nm<m<0.57 nm) (0.43 nm<m<0.45 nm)

B Ch. 19 Ch. 28 Ch. 13 (0.79 nm<m< 0.81nm) (1.50 nm<m<1.55 nm) (0.67 nm<m< 0.69 nm)

C Ch. 28(1.50 nm<m<1.55 nm)

D Ch. 93(8.21 nm<m<8.56 nm)

193


ral vegetation was created for each sector usingthe «natural environments» informative layer;this mask enabled us to improve the classifica-tion of vegetation classes as much as possibleand to remarkably shorten the processing time.Therefore, this selective approach does classifyareas of vegetation included in the natural envi-ronments, so removing urban areas, arablelands, etc.

In order to define correct vegetation classesthe naturalists of Rome Province carried outseveral phyto-sociological surveys in situ (in to-tal 160 field surveys performed on the Region-al Technical Map at scale 1:10000) accordingto Braun-Blanquet’s method that was further in-tegrated with previous works accomplished inthe surveyed area (Brandizzi et al., 1974; Ab-bate et al., 1981; Avena et al., 1990; Centro Re-

Table III. Definitions of prevailing vegetation classes.

Forest formations Shrubs formations Herbaceous formations Anthropics areas

Evergreen woods with prevalence Prevailing deciduous High mountain pastures Arable landsof Quercus ilex; Quercus suber and evergreen sclerophyll with Sesleria tenuifolia.woods with Styrax officinalis shrubs.and Castaneasativa (only for Catillo Mt. Natural Reserve).

Mixed woods with dominant Shrubs with Crataegus sp. Mountain and Permanent cropspresence of Quercus cerris; Prunus sp. Spartium sub-mountain pastures. (only for Natural mixed woods with local junceum, Cytisus sp.; Reserves)prevalence of Quercus robur shrubs with Erica arborea,(only for Nomentum and Cystus salvifolius Macchia di Gattaceca and Spartium junceumNatural Reserves). (only for Catillo

Mt. Natural Reserve).

Mixed woods with dominant Mountain and high Mesophylous Urban areaspresence of Quercus pubescens. and high mountain shrubs pastures.

with prevailing presenceJuniperus sp.

Mixed woods with dominant Thermophylous Ornamentalpresence of Ostrya carpinifolia. pastures with green

Ampelodesmos mauritanicus;garigues.

Castanea sativa woods. Bare surfaces; Quarriesrock outcrops.

Fagus sylvatica woods.

Riparian Vegetation with dominant Waterspresence of Populus albaand Salix alba.

Conifers woods with dominant Unclassifiedpresence Pinus nigra.

Thickets in evolution; thicketswith Cercis siliquastrum, Styrax officinalis, Pistacia terebinthus and Acer monspesulanum (only for Catillo Mt. Natural Reserve).

gionale per la Documentazione dei Beni Cultur-ali e Ambientali, 1993). The main physiognom-ic classes defined by the naturalist are reportedin table III. The in situ observations enabled usto identify spectrally homogeneous classes onthe airborne remotely sensed images, useful fordefining 133 Regions Of Interest (ROI, as im-plemented in the ENVI 3.4 software package;RSI, 2000), of which about 70% as training ar-eas, and 30% to further classification testing.The ROI were chosen to represent adequatelythe vegetation classes shown in table III, per-taining to forest, shrub and herbaceous forma-tions for a total of 17 vegetation classes; theywere selected within several morphologic con-texts to better represent the spectral class vari-ability.

The classification procedure is based on theMaximum Likelihood (ML) algorithm appliedto the first 28 bands (spectral range from 0.43-1.55 nm). The ML classifier was chosen be-cause allows a better recognition of those spec-tral classes, like the vegetation species, whichpresent very correlated spectra. This algorithmassumes that the statistics for each class in eachband are normally distributed and calculates theprobability that a given pixel belongs to a spe-cific class (Richards, 1999). Each pixel is as-signed to the class that shows the highest prob-ability.

The cartographic restitution of the final the-matic product was obtained by mosaiking thethree sectors into a GIS-compatible (GeoTIFF)image format, using the Gauss-Boaga East timezone reference system (fig. 4).

5. Results and discussion

The «Map of Prevailing vegetation classes»(fig. 4) also reproduces other classes pertaining tothe anthropic formations defined in table II. Thesurface percentages of the cover classes found thestudy area are as follows: forest formations35.10%, shrub formations 15.23%, herbaceousformations 14.18 %, bare surfaces and rock out-crops 0.65%, Waters bodies 0.11%, arable lands30.71%, urban areas 3.92%, quarries 0.11%; con-sequently the natural and semi-natural area cov-ers about 65% of the entire study area.

The classification accuracy of the thematicmap was estimated by confusion matrix and Kcoefficient (see table IV). Overall accuracy iscomputed by dividing the total correct pixels(sum of the major diagonal) by the total numberof pixels in the confusion matrix. Kappa analy-sis is a discrete multivariate technique used inaccuracy assessments, that is a measure ofagreement or accuracy (Congalton, 1991). Theevaluation procedure applied to the data gavean overall accuracy equal to 80.69% and a Kcoefficient equal to 0.7899.

The classification produced good results inthe discrimination of forest and shrub formationsand herbaceous formations. In particular, be-tween forest formations the methodology welldiscriminated conifers from broadleaves (e.g.,Pinaceae from Fagaceae), and between broad-leaves, evergreens from deciduous ones (e.g.,Quercus ilex woods from Castanea sativawoods), mainly in those areas where broad-leaves form pure and well developed forests.Less accurate results were obtained for the class-es that form mixed woods, among them: mixedwoods with dominant presence of Quercus cer-ris; mixed woods with dominant presence ofQuercus pubescens and mixed woods with dom-inant presence of Ostrya carpinifolia. The largevegetation biodiversity was highlighted locatingand mapping also vegetation classes with limitedextension and circumscribed within certain areas,or local prevailing species inside more extendedformations. From this point of view, the locationand mapping of Quercus suber woods and Cercissiliquastrum thickets, of high naturalistic and en-vironmental value, located in the Catillo Mt. Nat-ural Reserve, also represented an important re-sult. It was possible to localize Quercus roburcommunities within into Quercus cerris woods,present in the Nomentum and Macchia di Gat-taceca Natural Reserves. The identification of ar-eas with «anomalous» spectral responses (in thesector between the villages of S. Vito Romanoand Capranica Prenestina, Prenestini Mt.), inter-preted as Castanea sativa woods deteriorated byparasites’ attack, proved to be interesting as a fu-ture application hypothesis. Moreover it was pos-sible to map complex vegetation covers (mainlyshrubby formations), otherwise difficult to locatewith traditional techniques. The integrated ap-

194


195


Fig

.4.

Map

of

prev

ailin

g ve

geta

tion

clas

ses.

A –

Nat

ural

Res

erve

of

Nom

entu

m; B

– N

atur

al R

eser

ve o

f M

acch

ia d

i Gat

tace

ca; C

– R

egio

nal P

ark

of L

ucre

tili M

ts.;

D –

Reg

iona

l Par

k of

Sim

brui

ni M

ts.;

E –

Nat

ural

Res

erve

of

Cat

illo

Mt.

196


Table IV. Confusion matrix.

Ground Truth (pixels)

Confusion Sparse Crataegus sp., Deciduous, Rock Mesoph. Shrubs with Thermoph. Thickets Mountainmatrix veget. Spartium evergreen outcrops pastures Juniperus sp. pastures in evolution pastures

junceum, shrubsCytisus sp.

Class

Unclassif. 6 23 28 34 12 15 39 107 12

Sparse veg. 373 0 0 0 0 0 0 0 1

Crataegus sp., 0 125 5 0 0 9 0 59 0Spartiumjunceum,

Cytisus sp.

Deciduous, 0 0 280 0 1 12 0 27 19evergreen

shrubs

Rock 0 0 0 305 0 0 2 0 0outcrops

Mesoph. 0 2 0 0 442 17 0 0 11pastures

Shrubs with 0 3 2 0 6 258 0 0 16Juniperus sp.

Thermoph. 4 0 0 0 0 2 400 0 0pastures

Thickets 0 10 3 0 0 0 0 473 0in evolution

Mountain 0 0 6 0 7 11 0 1 533pastures

Quercus 0 2 3 0 0 0 0 8 0ilex

Quercus 0 0 0 0 0 0 0 0 0cerris

Ostrya 0 0 0 0 0 1 0 2 0carpinifolia

Castanea 0 0 0 0 1 0 0 4 0sativa

Fagus 0 0 0 0 0 0 0 0 0sylvatica

Quercus 0 0 8 0 2 6 0 9 0pubescens

Riparian veg. 0 3 0 0 0 1 0 0 0

Conifers 0 0 0 0 0 0 0 0 0

Total 383 168 335 339 469 331 441 687 592

proach between image processing and photoint-erpretation disclosed 22 land cover classes.

In particular, the analysis of the map showsthat in the western sector (to west of Tivoli city),

chiefly characterized by covers deriving from an-thropic activities, the areas devoted to sowingcultivations, woody cultivations (with a large dif-fusion of olive trees) and pastures prevail. It is al-

197


Quercus Quercus Ostrya Castanea Fagus Quercus Riparian Conifers Total Confusionilex cerris carpinifolia sativa sylvatica pubescens vegetation matrix

Class

34 101 164 26 39 67 1 4 710 Unclassif.

0 0 0 0 0 0 0 0 374 Sparse veg.

2 0 0 10 0 1 1 0 212 Crataegus sp.,Spartiumjunceum,

Cytisus sp.

1 1 8 0 0 12 0 1 360 Deciduous,evergreen

shrubs

0 0 0 0 0 0 0 0 307 Rockoutcrops

0 0 0 5 0 1 0 0 477 Mesoph.pastures

0 0 0 4 0 0 2 0 290 Shrubs withJuniperus sp.

0 0 0 0 0 0 1 0 407 Thermoph.pastures

13 1 2 2 0 4 0 0 508 Thicketsin evolution

0 0 0 0 1 0 0 0 559 Mountainpastures

451 0 17 1 0 70 6 18 575 Quercusilex

3 485 92 0 114 8 0 0 702 Quercuscerris

1 26 422 0 27 22 0 0 501 Ostryacarpinifolia

1 1 0 668 2 0 0 0 676 Castaneasativa

0 26 66 0 733 0 0 0 825 Fagussylvatica

87 10 81 4 0 327 0 5 538 Quercuspubescens

28 0 0 0 0 1 197 2 231 Riparian veg.

14 0 0 0 0 4 0 492 510 Conifers

636 649 853 720 916 515 208 521 8762 Total

so possible to note an evident building develop-ment of urban centers, mainly along the most im-portant roads, and a diffuse and progressive ur-banization of the surrounding countryside. The

most significant and extended natural areas arelocated in or very close to the Natural Reserves(Nomentum and Macchia di Gattaceca), last evi-dence of a habitat, once more extended, which

198


characterized the so called «agro romano»,namely the countryside around Rome.

The central and eastern sectors (to east ofTivoli city), mainly mountainous, are mostlycharacterized by relatively intact habitat, withwide continuous natural areas or in progress ofre-naturalization, extended and compact forestformations which are well developed, diffusedprairies and large shrubby formations. Shrubbyformations present a broader variety in theirfloristic composition and spatial distribution,forming, together with some pastures and sparseor grouped trees, vegetation mosaics character-ized by a great lack of cover homogeneity, andgiving the relevant land a wide variability of thedistributive pattern of herbaceous, shrubby andarboreal components. Such variability especial-ly occurs where agricultural activities are beingprogressively deserted (a phenomenon commonenough in this area) and uncultivated soil istending to re-naturalize. Anthropic activites aremainly concentrated along valleys (Aniene Val-ley) and plain areas.

6. Conclusions

The multidisciplinary formulation given tothe research activity planning, with the constantcontribution of expertise from different disci-plines, like botanists, naturalists and remotesensing experts, has certainly aided in the goodoutcome of the project.

The thematic maps produced have beenhelpful in the preliminary drawing up of theManagement Plans of the Natural Reserves,which should constitute the reference for thefuture land management policy and the en-hancement of biocompatible activities, in re-spect and safeguard of the environment.

The developed methodology enabled us torapidly and practically process MIVIS datacovering wide land extensions, and to attain themap of the main vegetation classes with a goodconfidence. The application of such procedurealso permitted us to produce maps with ancil-lary information of interest for the environmen-tal planners of the Province Administration.With reference to this the Rome Province Ad-ministration is completing the management and

environmental recovery plans of the areas Mt.Soratte, Mt. Guadagnolo, Travertini Acque Al-bule (Tivoli), pertaining in the area study, indi-cated to the European Community as area SICand ZPS.

The additional maps show other land covertypes, such as the different agricultural crops (e.g.,corn, wheat, beetroot, etc), artificial surfaces inurban areas (brick, asbestos-cement, etc.).

The MIVIS image processing proceduresapplied in this project are characterized by theuse of the correction of the atmospheric contri-bution due to the path radiance, a geometriccorrection procedure (including the land orog-raphy) and particularly for the application ofselective criteria between informative layers ofthe main land cover classes (first level classesin the classification hierarchy). This procedurethen enabled us to process many adjacentscenes in a fast and practical manner, to man-age classes statistically homogeneous, and as aconsequence to produce thematic maps at dif-ferent scales characterized by relatively high Kvalues.

The use of the above described proceduresallows us to develop a follow-on phase whichwill likely include classification refinements inareas presenting a strong dynamism, and thebetter diversification of some vegetation types(e.g., shrubby and herbaceous communities).The availability of a DEM will also permit us tocarry out a morphological study of the area withreference to the relations between morphologi-cal classes and vegetation classes, as alreadymade on the area of Mt. Soratte.

Based on the encouraging results obtainedso far by means of the airborne remotely senseddata on the area of Rome Province, interestingscenarios can be hypothesized for a future de-velopment of common activities (CNR - LocalAuthorities) finalized to plan and implement ofeffective actions for the safeguard and manage-ment of vegetation and forest resources.

Acknowledgements

The authors are grateful to Maurizio Poscol-ieri for valuable discussions and to Angela Mi-rabelli for help in improving the paper.

199


REFERENCES

ABBATE, G., G.C AVENA, C. BLASI and L. VERI (1981): Car-ta della vegetazione del M.te Soratte 1:10000, in Stu-dio delle Tipologie Fitosociologiche del Monte Soratte(Lazio) e loro Contributo nella Definizione Fito-geografica dei Complessi Vegetazionali Centro-Appen-ninici (P.F. «Promozione della qualità dell’ambiente»CNR, Roma), AQ/1/125, 1-41.

AVANZI, G., C.M. MARINO and S. PIGNATTI (1997): Geocod-ifica e mosaico su modello altimetrico digitale di im-magini MIVIS, in Atti della I Conferenza NazionaleASITA, p. 179.

AVANZI, G., A. PALOMBO and S. PIGNATTI (2006): MIVISimage geocoding experience on merging position atti-tude system data and public domain GPS stream (ASI-GeoDAF), Ann. Geophysics, 49 (1), 11-20 (this vol-ume).

AVENA, G.C., L. BONIFAZI, S. FASCETTI and L. RUBECA

(1990): Carta della Vegetazione del Territorio della IXComunita’Montana del Lazio, Compreso nei Limiti delParco dei M.ti Lucretili (scala 1:25000), (Dipartimen-to di Biologia Vegetale, Università degli Studi di Roma«La Sapienza»).

BERK, A., L. BERNSTEIN and D. ROBERTSON (1989): MOD-TRAN: a Moderate Resolution Model for LOW-TRAN7, Final Report, GL-TR-0122, AFGL, HanscomAFB, p. 42.

BLASI, C. (1994): Fitoclimatologia del Lazio, Fitosociol.,27, 151-175.

BORAK, J.S., E.F. LAMBIN and A.H. STRAHLER (2000): Theuse of temporal metrics for landcover change detectionat coarse spatial scales, Int. J. Remote Sensing, 21 (6-7), 1415-1432.

BRANDIZZI, G., A. ANGELUCCI, A .BIASINI, G. CASELLA, R.FUNICIELLO, M. GIANNINI, G. GIOVAGNOLI, L. RAMPI-CHINI, A. PANASCÈ and G. BIGI (1974): Carta Agrofore-stale della Provincia di Roma (Camera di Commercio,Industria, Artigianato e Agricoltura di Roma).

CAVALLI, R.M., L. FUSILLI, A. GUIDI, S. PANZARASA, S.PIG-NATTI, A. VINCI and C.M. MARINO (2001): Dati iper-

spettrali telerilevati per la gestione di riserve naturalidella Provincia di Roma, in Atti della V Conferenzanazionale ASITA, vol. I, 477-482.

CENTRO REGIONALE PER LA DOCUMENTAZIONE DEI BENI

CULTURALI ED AMBIENTALI (1993): Carta del Paesaggiovegetale della Valle del Tevere (scala 1:50 000), (Ass.toalla Cultura - Regione Lazio).

CONGALTON, R.G. (1991): A review of assessing the accura-cy of classifications of remotely sensed data, RemoteSensing Environ., 37, 35-46.

CONGALTON, R.G., K. GREEN and J. TEPLY (1993): Mappingold-growth forest on National Forests and park lands inthe Pacific Northwest from remotely sensed data, Pho-togramm. Eng. Remote Sensing, 59, 529-535.

GAO, B.-C. (1996): NDWI – A Normalized Difference Wa-ter Index for remote sensing of vegetation liquid waterfrom space, Remote Sensing Environ., 58, 257-266.

HILL, J., J. MEGIER and W. MEHL (1995): Land degradation,soil erosion, and desertification monitoring in Mediter-ranean ecosystems, Remote Sensing Rev., 12, 107-130.

LAMBIN, E.F. (1998): Modeling and monitoring land-coverchange pro-cesses in tropical Regions, Prog. Phys. Ge-ogr., 21 (3), 375-393.

MONTELUCCI, G. (1972): Considerazioni sul componenteorientale nelle foreste della Penisola, Ann. Accad. Ital.Sci. For., XXI, 121-169.

MONTELUCCI, G. (1978)(1976-1977): Lineamenti della veg-etazione nel Lazio, Ann. Bot. (Roma), XXXL-VI, 1-107.

MONTELUCCI, G. (1984): I monti di Tivoli dal punto di vistabotanico, Natura e Montagna, 3, 37-48.

MORENO, J.M. and W.C. OECHEL (Editors) (1995): GlobalChange and Mediterranean-Type Ecosystems (NewYork, Springer-Verlag), p. 527.

NEMANI, R. and S.W. RUNNING (1996): Global vegetationcover changes from coarse resolution satellite data, J.Geophys. Res., 101, 7157-7162.

RICHARDS, J.A. (1999): Remote Sensing Digital Image Ana-lysis (Springer-Verlag), p. 240.

RSI (2000): ENVI User’s Guide (Research Systems Inc.),261-288.

the natural areas of rome province detected by airborne ... · nies with large land holdings...

Documents