Advertisement

Landscape and Ecological Engineering

, Volume 15, Issue 2, pp 231–240 | Cite as

Using unmanned aerial vehicles for vegetation mapping and identification of botanical species in wetlands

  • Andrea BertacchiEmail author
  • Vittoria Giannini
  • Carmelo Di Franco
  • Nicola Silvestri
Report
  • 225 Downloads

Abstract

High-resolution aerial photographs have important applications in vegetation mapping, especially in environments, such as wetlands, which are not easily accessible by ground operators. Unmanned aerial vehicles (UAVs), equipped with cameras capable of taking photographs of < 1 cm pixel resolution, are promising not only for the vegetation mapping but also for the identification of plant species. This paper illustrated the results of three different flight heights (5 m = 3.5126 mm/pixel; 10 m = 7.0252 mm/pixel; 25 m = 17.5630 mm/pixel), using 12MP images and their magnification, on the identification of vegetation and botanical species in a rewetted peatland populated mainly by Phragmites australis and Myriophyllum aquaticum within the Massaciuccoli Lake basin (Northern Tuscany, Italy). Among the obtained images, we selected the best flight height for the vegetation mapping and the botanical identification of the plant species using both visual and automated image analyses. Images taken from flights at 25 m of height proved to be useful for a sufficiently detailed mapping, while those from 10 m of height were more suitable for the detection of plant microcommunities. However, the most accurate identification of the species (at the taxonomic level of genus/species) was possible only with the images taken from 5 m of height.

Keywords

UAVs Aerial photos Vegetation map Wetland Phragmites Myriophyllum Tuscany 

Introduction

Wetlands provide important ecosystem services, such as biodiversity and habitat conservation, flood protection (De Groot 1992; Carpenter and Cottingham 1997; Keddy 2000; Davies et al. 2008) and carbon sequestration (Zedler and Kercher 2005), from which we derive the importance of their conservation and restoration (Klimkowska et al. 2010; Verhoeven 2014). Moreover, the significant role of wetland plant species in water phytotreatment and the use of aquatic vegetation as an indicator of the environmental conditions of the aquatic systems have been widely studied (Vymazal 2007; Penning et al. 2008; Zak et al. 2014).

In this regard, monitoring wetland plant communities and hydro-hygrophilous flora plays a fundamental role in their management, conservation and/or restoration (Klötzli and Grootjans 2001; Oertli et al. 2002; Schmidt and Skidmore 2003; Yabe and Nakamura 2010).

Enhancing the detection of plants and plant communities in wetlands is a crucial information to assess the species occurrence and abundance at the species level. In particular, aerial photos could represent a good tool for vegetation mapping and the identification of plant communities in wetlands, thus improving field surveys (Howland 1980; Fitzgerald et al. 2005; Yichun et al. 2008; Fujimura et al. 2013). Moreover, the most recent systems of automated image classification can guarantee a consistent support (Adam et al. 2010). Nevertheless, until now, the availability of data at the appropriate spatial scale to detect some types of within-wetland community and botanical species has been limited (Henderson and Lewis 2008; Rebelo et al. 2009; Muro et al. 2016). Imageries from unmanned aerial vehicles (UAVs) are a relatively new source of data that can be profitably used in ecology or geobotany and, more broadly, in natural resource management (Watts et al. 2010; Shahbazi et al. 2014; Marcaccio et al. 2016). UAVs can provide imagery with the appropriate resolution (< 10 to < 2 cm pixel) not only for community level classification but also for the identification of a single botanical species (Getzin et al. 2012; Kaneko and Nohara 2014).

The perspective of using these tools becomes more feasible everyday from an economic perspective as the costs are decreasing with technological improvement (Anderson and Gaston 2013), although social impacts should be further investigated (Sandbrook 2015).

The purpose of this research was to identify the flight height and the image resolution fitting better with: (i) mapping a mature hydro-hygrophilous phytocoenoses and (ii) detecting botanical species otherwise invisible from the ground.

Materials and methods

The study was carried out in 2017 on a 2.7 ha rewetted peatland (Natural Wetland System—NWS) located in a larger phytotreatment pilot experimental (15 ha) set in Migliarino, San Rossore, Massaciuccoli Regional Park (Pisa, Italy—43°49′59.5″N; 10°19′50.7″E) (Ciccolini et al. 2013; Giannini et al. 2017) (Fig. 1).
Fig. 1

Geographical localization of investigated area

In particular, the pilot experimental field is located in a reclamation district of the coastal plain of the west-central Italy, characterized by large-scale, intensive agriculture and the presence of a vulnerable receiving water body constituted by the Massaciuccoli Lake. The natural vegetation landscape nearby the pilot experimental field is similar to that present in the surrounding of the Massaciuccoli Lake, where vegetation landscape is mainly formed by a mosaic of Phragmites australis (Cav.) Trin. ex Steud., Cladium mariscus (L.) Phol, Typha latifolia L. and T. angustifolia L. groupments (Phragminion communis Koch 1926), with patches of hydrophitic species such as Myriophyllum verticillatum L., Nymphea alba L., Lemna minor L. and a large spread of microwoods of hygrophilous phanerophytes such as Alnus glutinosa (L.) Gaertn, Frangula alnus Mill., and Salix sp.pl. (Alnion glutinosae Malcuit 1929, Salicion cinereae Müller and Görs 1958) (Bertacchi et al. 2015; Lastrucci et al. 2017).

The soils of the area can be classified as Histosol according to the USDA system and as Rheic Histosol according to the FAO system (Pellegrino et al. 2015). The climate is classified as Mediterranean (Csa) according to Köppen–Geiger climate classification map (Kottek et al. 2006). Summers are dry and hot, while rainfall is mainly concentrated in autumn and spring (mean annual rainfall = 910 mm), and mean air temperature at 2 m ranges from 6.6 to 21.8 °C (mean = 14.6) (http://www.sir.toscana.it/).

After rewetting, NWS evolved naturally, thus allowing the recolonization from spontaneous vegetation. The NWS was periodically monitored from 2013 to 2016 by ground surveys and aerial photos taken by a UAV, every summer (Giannini et al. 2018). In the final step of the experiment (spring–summer 2017), a UAV flying at 5, 10 and 25 m, with respective different flight paths, was used. The drone was an IRIS quadrotor controlled by a PX4 autopilot board and equipped with a GoPro camera set in a gimbal stabilizer. It was driven by four 850 kV brushless motors and powered by a 3S LiPo battery (11.1 V, 5.5 Ah). The total weight was approximately 1.3 kg. To compensate for rotation errors and keep the camera always aligned toward a specified direction, a gimbal stabilizer with two brushless motors was used. The GoPro camera was used at 12 MP (focal length = 2.92 mm, equivalent to 16.21 mm of 35 mm format; F-stop = f/2.8, Iso 100; 1/580 exposure; FOV = 122.6°) equipped by a panchromatic sensor. The PX4 board runs the open-source APM Ardupilot. In all the flights, the camera was set to shoot at every second. The 25-m flight took about 7 min with a flight speed of 7 m/s. The resulting overlap between two consecutive photos was 81%. The 10-m flight took about 15 min with a flight speed of 5 m/s. The resulting overlap between two consecutive photos was 66%. Finally, the 5-m flight took about 30 min (over 3 flights) with a flight speed of 5 m/s. The resulting overlap between two consecutive photos was 33%.

At the given altitudes, the obtained resolutions were, respectively, 5 m = 3.5126 mm/pixel; 10 m = 7.0252 mm/pixel and 25 m = 17.5630 mm/pixel.

Adobe Photoshop CC was used to remove distortion and correct the image back to a rectilinear lens projection, while Agistoft PhotoScan Professional Edition was used for mosaicking. Aerophotos were imported in GIS environment using Map Info® 10.1 for calibration, georeferencing and mapping. The orthomosaic of the 25-m flight has been created by composing 340 pictures, with a resulting reprojection error varying between 0.81 pixels and 8.43 pixels, with an average error equal to 2.47435 pixels.

For the vegetation mapping, an automated image classification based on object-based image analysis (OBIA) was tested (eCognition essential® software), jointly to the visual interpretation and the manual drawing in GIS Map Info.

To verify the effectiveness of the drone photos in the species identification, 24 frames taken, respectively, from flights at 25 m, 10 m, and 5 m of height along a transversal transect within the NWS have been selected. Obviously, the photos of the 5-m series corresponded to the original dimensions, while those of 10-m and 25-m series, which covered much larger areas, were cut at the same spatial dimensions as those at 5 m. The photos have been observed on the screen by three different botanists each of whom has been asked, by different frame and height, to identify the different species. At the same time, along the same transect and for each plot corresponding to the frame, the floristic census was carried out using both the ground survey and the survey by a boat.

With ground surveys, phytocenotic diversity was detected using the Braun-Blanquet (1983) method and this was compared with what was identified from aerial photos taken by a UAV. The images were interpreted assigning each digitized area to a specific phytocoenotic community. Flora has been randomly detected by field surveys, but also from UAV images that made possible the individuation of specimens that were previously hidden or unattainable. The floristic inventory was conducted over 5 ha, thus including also a part of the neighbouring land while the vegetation mapping was carried out exclusively within the NWS.

Results and discussion

Although the image elaboration proved to be much faster and more accurate than the human operator on the edge tracing, it revealed a low accuracy in identifying the different phytocoenoses and, even more so, the single specimen. In addition, the long time required for the image processing (e.g., segmentation settings, merging, etc.) was not functional for the purpose of this study. Thus, the relatively small surface of the investigated area and the need of a high accuracy in the detection drove toward the visual analysis and the vegetation mapping by manual drawing in GIS Map Info.

The accurate visual analysis of the photos taken from 25 m of height revealed the ability to locate the different phytocenoses in the site with good accuracy. Furthermore, these frames had the undoubtable benefit of faster execution and assembly times, and in much larger areas, they may prove to be better utilized.

The 25-m flight photos (17.5630 mm/pixel) were chosen for the orthophotomosaic construction and for the subsequent realization of the vegetation map (Fig. 2). This flight height allowed a high-pixel definition for mapping (original scale 1:1000) proved by the well-perceptible difference among the physiognomy and the chromatic texture of the different plant communities. Nevertheless, the magnification of the frames up to the maximum pixel resolution did not allow us to go further than the main phytocenosis. The visual analysis of the 10-m flight frames (10 m = 7.0252 mm/pixel) and their magnification was also able to provide the exact identification and localization of some plant microcommunities/single specimen scattered in the mosaic. The 5-m flight photos (5 m = 3.5126 mm/pixel) showed big problems related to the frame stitching because of the considerable distortion at the sides of the frame and an excessively high number of photos for mosaicking.
Fig. 2

Orthophotographic map of vegetation

The vegetation surveys by the ground, based on the differences identified by the aerophotos, revealed the presence of four floristically and physiognomically distinct plant communities (Table 1). From the phytosociological point of view, two plant communities can be ascribed to Phragmition communis Koch 1926 (surveys no: 8, 10, 12, 25, 27), Nymphaeion albae Oberdorfer 1957 (surveys no: 28, 30) and Chenopodion rubri (Tüxen 1960) Hilbig & Jage 1972 (surveys no: 31, 32) (Biondi et al. 2014; http://www.prodromo-vegetazione-italia.org/). Vegetation mapping showed a great extension of two main plant populations: Phragmites australis communities (1.4 ha) and Myriophyllum aquaticum communities (0.78 ha). Differently on a minor surface, there were Typha latifolia populations (0.2 ha), while in temporary emerged areas and those immediately outside the NWS, prevalent humid Echinochloa crus-galli grasslands predominated (0.16 ha) (Fig. 2). The identified flora indicates a clear prevalence of hydro-hygrophilous species, native and common in the humid areas of the surrounding area, except for the alien and invasive Myriophyllum aquaticum of South-American origin (Bertacchi et al. 2015). Of the 66 species found overall in the area, 5 species have been detected, thanks to the analysis of the 5-m flight, showing the specimen and the precise detection point otherwise invisible from the ground (Table 2).
Table 1

Table of the most significant surveys of the plant communities detected

Survey no

8

10

12

25

27

28

30

31

32

Surface (sq)

25

25

25

9

4

60

100

25

25

Coverage (%)

100

100

100

80

100

80

100

50

70

Species no

6

6

7

4

3

4

2

11

13

 

PH

TY

MY

EC

Phragmites australis (Cav.) Trin.

5

4

5

1

.

.

.

1

+

Calystegia sepium (L.) R.Br.

+

+

+

+

1

.

.

+

+

Eupatorium cannabinum L.

.

+

+

.

.

.

.

.

r

Stachys palustris L.

+

+

+

.

.

.

.

.

.

Lythrum salicaria L.

.

+

+

+

.

+

.

+

+

Typha latifolia L.

r

.

.

3

4

.

+

.

.

Schoenoplectus tabernaemontani (Gmel.) Palla

.

.

r

.

.

.

.

+

r

Mentha aquatica L.

r

.

.

.

.

.

.

.

.

Iris pseudacorus L.

.

+

+

.

.

.

.

.

.

Oenanthe aquatica L.

+

.

.

.

.

.

.

.

.

Apium nudiflorum (L.) Lag

.

.

.

.

+

.

.

.

.

Myriophyllum aquaticum (Vell). Verdc.

.

.

.

.

.

5

5

.

.

Juncus articulatus L.

.

.

.

.

.

+

.

.

.

Lemna minor L.

.

.

.

.

.

+

.

.

.

Echinochloa crus-galli (L.) P. Beauv.

.

.

.

.

.

.

.

3

3

Poa trivialis L.

.

.

.

.

.

.

.

+

.

Juncus effusus L.

.

.

.

.

.

.

.

r

.

Juncus bufonius L.

.

.

.

.

.

.

.

.

+

Carex otrubae Podp.

.

.

.

.

.

.

.

r

r

Iris pseudacorus L.

.

.

.

.

.

.

.

.

r

Paspalum dilatatum Poir.

.

.

.

.

.

.

.

+

+

Ranunculus scleratus L.

.

.

.

.

.

.

.

.

+

Ranunculus sardous Crantz

.

.

.

.

.

.

.

+

+

Samolus valerandi L.

.

.

.

.

.

.

.

+

.

Epilobium hirsutum L.

.

.

.

.

.

.

.

.

+

Index values according to abundance/dominance Braun-Blanquet (1983) scale

PH Phragmites australis communities, TY Typha latifolia communities, MY Myriophyllum aquaticum communities, EC Echinocloa crus-galli communities

Table 2

Cheklist of identified plant species

Family

Species

Area

Alismataceae

Alisma plantago-aquatica L.A

a

Amaranthaceae

Amaranthus retroflexus L.

b

 

Chenopodium album L.

b

Araceae

Lemna minor L.

a

Apiaceae

Apium nodiflorum (L.) Lag

a

 

Daucus carota L.

b

 

Torilis arvensis (Huds.) Link subsp. arvensis

b

Asteraceae

Artemisia verlotorum Lamotte

ab

 

Bidens tripartita L.

ab

 

Cirsium arvense (L.) Scop.

b

 

Cichorium intybus L.

b

 

Crepis vesicaria L.

b

 

Eupatorium cannabinum L.

a

 

Matricaria camomilla L.

b

 

Picris hieracioides L.

b

 

Sonchus arvensis L.

b

 

Symphyotrichum squamatum (Spreng.) G. L. Nesom

b

 

Xanthium strumarium L.

ab

Boraginaceae

Echium vulgare L.

b

Brassicaceae

Capsella bursa-pastoris (L.) Medicus

b

Caprifoliaceae

Sambucus ebulus L.

ab

Caryophyllaceae

Silene alba (Miller) Krause

b

 

Stellaria media (L.) Vill.

b

Convolvulaceae

Calystegia sepium (L.) R.Br.A

ab

 

Convolvulus arvensis L.

b

Cyperaceae

Carex otrubae Podp.

ab

 

Cyperus glomeratus L.

a

 

Cyperus longus L.

a

 

Schoenoplectus tabernaemontani (Gmel.) Palla

a

Equisetaceae

Equisetum palustre L.

ab

Haloragaceae

Myriophyllum aquaticum (Vell.) Verdc

a

Juncaceae

Juncus acutus L.

a

 

Juncus articulatus L.

ab

 

Juncus bufonius L.

ab

 

Juncus effusus L.

ab

Lamiaceae

Mentha aquatica L.A

ab

 

Lycopus europaeus L.A

a

Fabaceae

Lotus corniculatus L.

b

 

Trifolium hybridum L.

b

 

Trifolium repens L.

b

Lythraceae

Lythrum salicaria L.

a

Onagraceae

Epilobium hirsutum L.

ab

Phytolaccaceae

Phytolacca americana L.

ab

Plantaginaceae

Plantago major L.

b

Poaceae

Arundo donax L.

a

 

Cynodon dactylon (L.) Pers.

ab

 

Echinochloa crus-galli (L.) P. Beauv.

ab

 

Holcus lanatus L.

a

 

Lolium perenne L.

a

 

Paspalum dilatatum Poir.

ab

 

Phragmites australis (Cav.) Trin.

ab

 

Poa trivialis L.

ab

 

Polypogon monspeliensis (L.) Desf.A

ab

 

Polypogon viridis (Gouan) Breistr.

ab

 

Setaria viridis L.

ab

Polygonaceae

Polygonum persicaria L.

b

 

Rumex conglomeratus Murray

b

Primulaceae

Samolus valerandi L.

a

Ranunculaceae

Ranunculus sardous Crantz

ab

 

Ranunculus sceleratus L.

ab

Rosaceae

Rubus ulmifolius Schott.

b

Scrophulariaceae

Linaria vulgaris Milerr

b

 

Verbascum pulverulentum Vill.

b

Solanaceae

Datura stramonium L.

b

Typhaceae

Typha latifolia L.

a

Urticaceae

Urtica dioica L.

b

a pool area, b outside area

ASpecies detected thanks to the drone

The detection of the single species by the photo interpretation gave different results among the three different heights. The number of species identified by each of the three observers did not appear to be significantly different for the series of photos, as the difference related to the possibility of identification between 25-m flight photos and 10-m flight photos was not significant. On the contrary, about the 5-m flight photos, the identification of the species appeared significantly higher and, moreover, very close to the actual number of species found on the ground. It was also significant that with few species, there was no substantial difference among the frames from different flight heights. As the number of species increased, the flight at 5 m became significantly discriminating on the species identification (Fig. 3).
Fig. 3

Number of species identified by the three photo interpreters by plots and by the three different flight heights and species actually present (error bars = stand. err.)

In the case of larger areas and for a simple physiognomic identification of plant communities, the 25 m flying altitude can be useful and sufficient. In this experience, on a relatively small surface, flight altitude of 25 m with a 12 MP of resolution could quickly solve the interpretation and the mapping of vegetation communities, with a relatively small number (340) of frames. Macrocommunities like those at Phragmites australis and Myriophyllum aquaticum were also easily identifiable and spatially quantifiable. In other cases, the detailed detection of different vegetation types and the identification of small plant communities or species mixed in the vegetation mosaic were more accurate using the frames taken from the 10-m flight. Zooming the photo, the species not distinguishable with the 25-m flight photos were partially detectable with the 10-m flight photos, although only with 5-m flight photos, the species were better identifiable by the botanical and taxonomical point of view (Fig. 4). Thus, the discovery of species located in spots not visible from the ground such as Alisma plantago-aquatica, Lycopus europaeus, Polypogon monspeliensis, and Mentha aquatica as well as the presence of specimens of Calystegia sepium hidden between reedbeds (Phragmites australis) was possible (Fig. 5).
Fig. 4

Examples of photos at 25-, 10-, and 5-m flight heights (ac) and the corresponding details at maximum resolution (a 25 m = 17.5630 mm/pixel; b 10 m = 7.0252 mm/pixel; c 5 m = 3.5126 mm/pixel)

Fig. 5

Species identification from 5-m flight height: a: Alisma plantago-aquatica; b: Typha angustifolia; c: Lycopus europaeus; d: Myriophyllum aquaticum; e: Lythrum salicaria; f: Phragmites australis; g: Polypogon monspeliensis; h: Xanthium strumarium; i: Echinochloa crus-galli; Calystegia sepium specimens (white dots)

Remote sensing based on the use of UAVs is a fast, non-intrusive and supplementary method to ground-based approaches for mapping vegetation of wetlands, swamps, and bogs (Klemas 2013). Although with drones it is possible to identify very small objects even from higher flight altitudes than those we used (Martin et al. 2012), the detection and identification of herbaceous species grown in areas not visible from the ground require a higher spatial resolution, even lower than 1 cm (Ishihama et al. 2012). From the results we obtained, the flight at 5 m can be occasionally used in microareas previously identified as an integration for the vegetation interpretation and the individuation of unclear specimens. However, often at this extremely low height, the UAV showed a higher risk of crash against the fronds of shrubs/trees. Therefore, by these results, drone images can be extremely useful and exhaustive for mapping the main vegetation components and coverage into wetlands at 25- to 10-m flight altitude.

In cases of limited surfaces and for precise botanical classification, the photo interpretation by an expert human operator seemed preferable to those by image processing. Although automated analysis is a robust way to deal with image heterogeneity, until now it seems to be more functional for large homogenous areas, like rangelands or sparse forest or, more generically, landform (Bryson et al. 2010; Blaschke 2010; Laliberte and Rango 2011); while, in environments physiognomically complex like wetlands, visual analysis is still preferable (Husson et al. 2014). The uniform chromatism resulting from the panchromatic images of these types of habitats often makes the pre-treatment of the image for the analysis process excessively long and laborious. In addition, often differences in illumination (e.g., sun angle, water glitter, shadows) can determine gross mistake and high misclassification by automated analysis. The photo interpreter uses both colour and form differences among the leaves’ shape and the plant structure for the individuation of the different species. Anyhow this choice is justified by the necessity of a botanist expert in field investigation and in the determination of botanical species. Therefore, ultimately, in our opinion, the unmanned aerial system, and more generally remote sensing, should be better interpreted only as a powerful integrative but not as exhaustive methods of investigation in the floristic and phytosociological studies, where the exact identification of species or subspecies is required and, in any case, field survey is always necessary.

Notes

Acknowledgements

This work was supported by the “Consorzio di Bonifica Versilia—Massaciuccoli” later “Consorzio di Bonifica 1 Toscana Nord” and funded by the “Regione Toscana”.

References

  1. Adam E, Mutanga O, Rugege D (2010) Multispectral and hyperspectral remote sensing for identification and mapping of wetland vegetation: a review. Wetl Ecol Manag 18:28–296CrossRefGoogle Scholar
  2. Anderson K, Gaston JK (2013) Lightweight unmanned aerial vehicles will revolutionize spatial ecology. Front Ecol Environ 113:138–146.  https://doi.org/10.1890/120150 CrossRefGoogle Scholar
  3. Bertacchi A, Lombardi T, Tomei P, Labate M, Fontanelli A, Picchi L (2015) Vegetation landscape management of “Natural reserve of Chiarone”, Massaciuccoli lake basin Tuscany, IT. In: 2nd international congress of plant sciences, 110th congress of Italian Society of Botany, Pavia. http://www.societabotanicaitaliana.it/110/abstract_2015_congresso.pdf. Accessed 15 Jan 2018
  4. Biondi E, Blasi C, Allegrezza M, Anzellotti I, Azzella MM, Carli E et al (2014) Plant communities of Italy: the vegetation prodrome. Plant Biosyst 1484:728–814.  https://doi.org/10.1080/11263504.2014.948527 CrossRefGoogle Scholar
  5. Blaschke T (2010) Object based image analysis for remote sensing. J Photogramm Remote Sens 65:2–16CrossRefGoogle Scholar
  6. Braun-Blanquet J (1983) Plant sociology: the study of plant communities. Koeltz Scientific Books, Koenigstein im TaunusGoogle Scholar
  7. Bryson M, Reid A, Ramos F, Sukkarieh S (2010) Airborne vision-based mapping and classification of large farmland environments. J Field Robot 27:632–655CrossRefGoogle Scholar
  8. Carpenter SR, Cottingham KL (1997) Resilience and restoration of lakes. Conservation ecology. Advance online publication. http://www.consecol.org/vol1/iss1/art2/ Accessed 15 Sept 2017
  9. Ciccolini V, Giannini V, Pistocchi C, Bosco S, Pellegrino E, Sabbatini, et al (2013)Restoration of a Mediterranean drained peatland: the case study of the Massaciuccoli lake basin Tuscany IT. AWARE: approaches in wetland restoration—focus on fen landscapes, Warsaw. http://www.biol.uw.edu.pl/aware/AWARE%20posters/AWARE_extended_abstract_Ciccolini%2020%2011%202013.pdf. Accessed 20 June 2018
  10. Davies B, Biggs J, Williams P, Whitfield M, Nicolet P, Sear D, Bray S, Maund S (2008) Comparative biodiversity of aquatic habitats in the European agricultural landscape. Agric Ecosyst Environ 125:1–8CrossRefGoogle Scholar
  11. De Groot RS (1992) Functions of nature: evaluation of nature in environmental planning, management and decision making. Wolters-Noordhoff, GroningenGoogle Scholar
  12. Fitzgerald DG, Zhu B, Hoskins SB, Haddad DE, Green KN, Rudstam LG, Mills EL (2005) Quantifying submerged aquatic vegetation using aerial photograph interpretation. Fisheries 30(2):61–73.  https://doi.org/10.1577/1548-8446(2005)30%5b61:QSAVUA%5d2.0.CO;2 CrossRefGoogle Scholar
  13. Fujimura Y, Takada M, Fujita H, Inoue T (2013) Change in distribution of the vascular plant Sasa palmata in Sarobetsu Mire between 1977 and 2003. Landsc Ecol Eng 9:305–309.  https://doi.org/10.1007/s11355-012-0193-4 CrossRefGoogle Scholar
  14. Getzin S, Wiegand K, Schöning I (2012) Assessing biodiversity in forests using very high-resolution images and unmanned aerial vehicles. Methods Ecol Evol 3:397–404.  https://doi.org/10.1111/j.2041-210X.2011.00158.x CrossRefGoogle Scholar
  15. Giannini V, Silvestri N, Dragoni F, Pistocchi C, Sabbatini T, Bonari E (2017) Growth and nutrient uptake of perennial crops in a paludicultural approach in a drained Mediterranean peatland. Ecol Eng 103:478–487CrossRefGoogle Scholar
  16. Giannini V, Bertacchi A, Bonari E, Silvestri N (2018) Rewetting of a former drained Mediterranean peatland in Tuscany for phyto-treatment use. J Environ Manag 208:92–101CrossRefGoogle Scholar
  17. Henderson FM, Lewis AJ (2008) Radar detection of wetland ecosystems: a review. Int J Remote Sens 29:5809–5835.  https://doi.org/10.1080/01431160801958405 CrossRefGoogle Scholar
  18. Howland WG (1980) Multispectral aerial photography for wetland vegetation mapping. Photogramm Eng Remote Sens 46:87–99Google Scholar
  19. Husson E, Hagner O, Ecke F (2014) Unmanned aircraft systems help to map aquatic vegetation. Appl Veg Sci 17:567–577CrossRefGoogle Scholar
  20. Ishihama F, Watabe Y, Oguma H (2012) Validation of a high-resolution, remotely operated aerial remote-sensing system for the identification of herbaceous plant species. Appl Veg Sci 15:383–389.  https://doi.org/10.1111/j.1654-109X.2012.01184.x CrossRefGoogle Scholar
  21. Kaneko K, Nohara S (2014) Review of effective vegetation mapping using the UAV unmanned aerial vehicle method. J Geogr Inf Sys 6:733–742.  https://doi.org/10.4236/jgis.2014.66060 CrossRefGoogle Scholar
  22. Keddy PA (2000) Wetland ecology: principles and conservation. Cambridge University Press, CambridgeGoogle Scholar
  23. Klemas V (2013) Remote sensing of emergent and submerged wetlands: an overview. Int J Remote Sens 34:6286–6320.  https://doi.org/10.1080/01431161.2013.800656 CrossRefGoogle Scholar
  24. Klimkowska A, DiggelenR Van, Grootjans AP, Kotowski W (2010) Prospects for fen meadow restoration on severely degraded fens. Perspect Plant Ecol Evolut Syst 12:245–255CrossRefGoogle Scholar
  25. Klötzli F, Grootjans AP (2001) Restoration of natural and semi-natural wetland systems in Central Europe: progress and predictability of developments. Restor Ecol 9:209–219.  https://doi.org/10.1046/j.1526-100x.2001.009002209.x CrossRefGoogle Scholar
  26. Kottek M, Grieser J, Beck C, Rudolf B, Rubel F (2006) World Map of the Köppen-Geiger climate classification updated. Meteorol Z 153:259–263CrossRefGoogle Scholar
  27. Laliberte AS, Rango A (2011) Image processing and classification procedures for analysis of sub-decimeter imagery acquired with an unmanned aircraft over arid rangelands. GISci Remote Sens 48:4–23CrossRefGoogle Scholar
  28. Lastrucci L, Dell’Olmo L, Foggi B, Massi L, Nuccio C, Vicenti C, Viciani D (2017) Contribution to the knowledge of the vegetation of the Lake Massaciuccoli northern Tuscany, Italy. Plant Sociol 54:67–87.  https://doi.org/10.7338/pls2017541/03 CrossRefGoogle Scholar
  29. Marcaccio JV, Markle CE, Chow-Fraser P (2016) Use of fixed-wing and multi-rotor unmanned aerial vehicles to map dynamic changes in a freshwater marsh. J Unmanned Veh Syst 43:193–202.  https://doi.org/10.1139/juvs-2015-0016 CrossRefGoogle Scholar
  30. Martin J, Edwards HH, Burgess MA, Percival HF, Fagan DE et al (2012) Estimating distribution of hidden objects with drones: from tennis balls to Manatees. PLoS One 76:e38882.  https://doi.org/10.1371/journal.pone.0038882 CrossRefGoogle Scholar
  31. Muro J, Canty M, Conradsen K, Hüttich C, Nielsen AA et al (2016) Short-term change detection in wetlands using sentinel-1 time series. Remote Sens 8:795.  https://doi.org/10.3390/rs8100795 CrossRefGoogle Scholar
  32. Oertli B, Joyce DA, Castella E, Juge R, Cambin D, Lachavanne JB (2002) Does size matter? The relationship between ponds area and biodiversity. Biol Conserv 104:59–70CrossRefGoogle Scholar
  33. Pellegrino E, Bosco S, Ciccolini V, Pistocchi C, Sabbatini T, Silvestri N, Bonari E (2015) Agricultural abandonment in Mediterranean reclaimed peaty soils: long-term effects on soil chemical properties, arbuscular mycorrhizas and CO2 flux. Agric Ecosyst Environ 199:164–175CrossRefGoogle Scholar
  34. Penning WE, Dudley B, Mjelde M, Hellsten S, Hanganu J, Kolada A, van den Berg M, Poikane S et al (2008) Using aquatic macrophyte community indices to define the ecological status of European lakes. Aquat Ecol 42:253–264CrossRefGoogle Scholar
  35. Rebelo LM, Finlayson CM, Nagabhatla N (2009) Remote sensing and GIS for wetland inventory, mapping and change analysis. J Environ Manag 90:2144–2153CrossRefGoogle Scholar
  36. Sandbrook C (2015) The social implications of using drones for biodiversity conservation. Ambio 444:s636–s647.  https://doi.org/10.1007/s13280-015-0714-0 CrossRefGoogle Scholar
  37. Schmidt K, Skidmore A (2003) Spectral discrimination of vegetation types in a coastal wetland. Remote Sens Environ 85:92–108.  https://doi.org/10.1016/S0034-425702.00196-7 CrossRefGoogle Scholar
  38. Shahbazi M, Théau J, Ménard P (2014) Recent applications of unmanned aerial imagery in natural resource management. GIScience Remote Sens 514:339–365.  https://doi.org/10.1080/15481603.2014.926650 CrossRefGoogle Scholar
  39. Verhoeven JTA (2014) Wetlands in Europe: perspective for restoration of a lost paradise. Ecol Eng 66:6–9CrossRefGoogle Scholar
  40. Vymazal J (2007) Removal of nutrients in various types of 645 constructed wetlands. Sci Total Environ 380:48–65CrossRefPubMedGoogle Scholar
  41. Watts AC, Perry JH, Smith SE, Burgess MA, Wilkinson BE, Szantoi Z, Ifju PG, Percival HF (2010) Small unmanned aircraft systems for low-altitude aerial surveys. J Wildl Manag 747:1614–1619.  https://doi.org/10.2193/2009-425 CrossRefGoogle Scholar
  42. Yabe K, Nakamura T (2010) Assessment of flora, plant communities, and hydrochemical conditions for adaptive management of a small artificial wetland made in a park of a cool-temperate city. Landsc Ecol Eng 6:201–210.  https://doi.org/10.1007/s11355-009-0098-z CrossRefGoogle Scholar
  43. Yichun X, Zongyao S, Mei Y (2008) Remote sensing imagery in vegetation mapping: a review. J Plant Ecol 1:9–23.  https://doi.org/10.1093/jpe/rtm005 CrossRefGoogle Scholar
  44. Zak D, Gelbrecht J, Zerbe S, Shatwell T, Barth M, Cabezas A, Steffenhagen P (2014) How helophytes influence the phosphorus cycle in degraded inundated peat soils—implications for fen restoration. Ecol Eng 66:82–90CrossRefGoogle Scholar
  45. Zedler JB, Kercher S (2005) Wetland resources: status, trends, ecosystem services, and restorability. Annu Rev Environ Resour 30:39–74CrossRefGoogle Scholar

Copyright information

© International Consortium of Landscape and Ecological Engineering and Springer Japan KK, part of Springer Nature 2019

Authors and Affiliations

  1. 1.Department of Agriculture, Food and Environment (DAFE)University of PisaPisaItaly
  2. 2.Institute of Life Sciences, Scuola Superiore Sant’AnnaPisaItaly
  3. 3.TECIP Scuola Superiore Sant’AnnaPisaItaly

Personalised recommendations