International Journal of Applied Earth Observation and Geoinformation
Broad-scale spatial pattern of forest landscape types in the Guiana Shield
Research highlights
► We identity five remotely sensed landscape classes (RSLC) in Guiana Shield forests. ► Forest RSLCs differ by their canopy openness. ► Distributions of forest RSLCs mainly fit with climatic gradients. ► The Guianan dense forest arch is the most telling pattern revealed by our study.
Introduction
At a time when Amazonian tropical rainforests are undergoing profound and rapid changes (Laurance et al., 2001), characterizing their spatial organisation is of major importance for analysing land-use changes and promoting sustainable management of forest resources. From a conservation point of view, the question arises as to how far large protected areas actually encompass the diversity of forest ecosystems. Landscape is a key concept at a key scale for sustainable management, but is not fully taken into consideration by ecologists and managers.
A landscape is an area that is spatially heterogeneous in at least one factor of interest (Turner, 2005). Consequently, landscape elements may be defined as homogeneous areas with respect to a set of factors of interest. Following Forman and Godron (1986), we define landscape types as a recognisable combinations and proportions of landscape elements. The mapping of such landscape types across large areas allow to detect what we call here the spatial pattern of landscape types. Our study focuses on the forest landscapes types of the Guiana Shield area.
At local scale, many authors have described an impressive number of landscape elements with specific typologies according to their structure (i.e. density and size-frequency distribution of stems, canopy structure, and floristics) either from field observations (Pires and Prance, 1985), quantitative measurements (Clark and Clark, 2000) or from high or very high spatial resolution satellite imagery (Lu et al., 2003, Johansen et al., 2007). All these studies involved a context-dependent typology and a fine scale resolution, both inaccurate for the mapping of large trends in the variations of vegetation cover.
At regional scale, maps of forest landscape types have been based on broad environmental classes rather than on actual vegetation characteristics. The pioneering RADAMBRASIL and its improvements (Veloso et al., 1991, IBGE, 1992) was the first attempt to map the spatial distribution of native landscape types in Brazil, mostly based on proxies of their determinant factors such as altitudinal classes. Similar approaches have been used in the Guiana Shield countries (Guyana: ter Steege, 2001, Huber et al., 1995; Suriname: Ravillous, 2000; French Guiana: Girou, 2001), combining in so different ways high resolution remote sensing (typically 30 m spatial resolution from Landsat imagery) and field data that they are mostly not compatible and unable to produce a global view of the spatial pattern of forest landscape types.
At continental scale, according to remote sensing, the forest is clearly distinguished from other biomes (e.g. savannas, wetlands, shrubland, and agriculture; Friedl et al., 2002, Bartholomé et al., 2004) but the resulting vegetation maps display tropical rainforests as a uniform broad-leaved evergreen forest class and fail to distinguish any distinctive forest landscape type within (South America: Eva et al., 2004; Asia: Stibig et al., 2007; Africa: Mayaux et al., 2004). On the contrary, studies based on scattered, but detailed field measurements identified clear patterns of species diversity, (ter Steege et al., 2000), life traits (ter Steege et al., 2006), or aboveground biomass (Malhi et al., 2006, Saatchi et al., 2007), thus strongly suggesting the existence of such distinctive types. Therefore, depicting such geographical structure in the forest vegetation cover from continuous data is of major interest, because it would provide key elements to support forest conservation and management policies. The objective of this study is to identify, characterize and map distinct forest landscape types within the evergreen lowland rainforest at the sub-continental scale of the Guiana Shield.
Section snippets
Location
The Guiana Shield covers 2.3 million km2 and is located on the north-eastern part of continental South America. The study area lies between latitude 11° North and 4° South, and between longitude 48° and 58° West. The climate is tropical with annual precipitation ranging from 1500 to 4000 mm and mean annual temperatures ranging from 25 °C to 30 °C (Hammond, 2005). The number of consecutive months with less than 100 mm precipitation varies from 0 to 7 (Sombroek, 2001). The altitude ranges from 0 to 3000
Results
The spectral classification yielded 33 remotely sensed landscape classes (RSLC) hereafter numbered from the lowest (RSLC-1) to the highest (RSLC-33) reflectance values (Table 3). Twelve main groups of RSLC were identified (Fig. 3) corresponding to the main landscape types. Among these, RSLC-2–13 correspond to the wide variety of continental aquatic environments and their associated landscape types, to which periodically flooded open areas (RSLC-27, -29, -31 and -33) are spatially linked. Four
Discussion
The main result of our study is to show, for the first time, a large scale spatial pattern of five distinct forest landscape types within the evergreen lowland forests of the Guiana Shield, using remote sensing data. This result greatly improves our vision of tropical forests because so far, at continental scale, lowland tropical rainforests were either mapped as a single homogeneous class (namely tropical lowland forest or broadleaved evergreen tropical forest) (Eva et al., 2004 for Amazonia,
Conclusion
The main interest of this study was to provide a homogeneous view of forest landscape types and to reveal a broad spatial pattern within the tropical evergreen forest at the scale of the Guiana Shield. This spatial pattern appears to be linked to climatic conditions. Identifying spatial pattern of forest landscape types is crucial for the improvement of management and conservation strategies (Bush and Lovejoy, 2007). The ability to identify spatial pattern could significantly improve the carbon
Acknowledgements
The authors would like to thank the Centre National d’Etudes Spatiales (CNES) and the European Commission Joint Research Centre for allowing us access to Spot-4/VEGETATION data sets. Thanks to the National Aeronautics and Space Administration (NASA) for providing the Shuttle Radar Terra Mission (SRTM) data set. This study was funded by the 12th Contrat de Plan Etat-Région of French Guiana (CAREFOR project) and by the French Overseas Ministry. All participants are thanked for their contributions
References (54)
- et al.
Paleo-environmental change in Amazonian and African rainforest during the LGM
Palaeogeography, Palaeoclimatology, Palaeoecology
(2006) - et al.
Assessing the extent of agriculture/pasture and secondary succession forest in the Brazilian Legal Amazon using SPOT VEGETATION data
Remote Sensing of Environment
(2006) - et al.
Landscape-scale variation in forest structure and biomass in a tropical rain forest
Forest Ecology and Management
(2000) - et al.
Global land cover mapping from MODIS: algorithms and early results
Remote Sensing of Environment
(2002) - et al.
Application of high spatial resolution satellite imagery for riparian and forest ecosystem classification
Remote Sensing of Environment
(2007) - et al.
Linear mixture model applied to Amazonian vegetation classification
Remote Sensing of Environment
(2003) - et al.
Analysis of lacunarity and scales of spatial homogeneity in IKONOS images of Amazonian tropical forest canopies
Remote Sensing of Environment
(2008) - et al.
Mapping and characterizing the vegetation types of the Democratic Republic of Congo using SPOT VEGETATION time series
International Journal of Applied Earth Observation and Geoinformation
(2009) - et al.
Secondary forest detection in a Neotropical dry forest using Landsat 7 ETM+ imagery
Biotropica
(2005) - et al.
Global land cover for the year 2000, land cover classification produced with data acquired in 2000 from the VEGETATION instrument, onboard the SPOT-4 satellite, 1/25.500.000 scale map
Impact of severe dry season on net ecosystem exchange in the neotropical rainforest of French Guiana
Global Change Biology
Carte géomorphologique de la Guyane
Amazonian conservation: pushing the limits of biogeographical knowledge
Journal of Biogeography
Predicting tropical forest stand structure parameters from Fourier transform of very high-resolution remotely sensed canopy images
Journal of Applied Ecology
Carte géologique de la France à 1:500 000
A land cover map of South America
Global Change Biology
Species richness and floristic composition in four hectares in the Jau national Park in upland forests in Central Amazonia
Biodiversity and Conservation
Landscape Ecology
Paysages et végétation, Atlas illustré de la Guyane, planche n°16, 1:1,300,000°
Ecology and management of a neotropical forest
Lessons Drawn from Paracou, a Long-term Experimental Research site in French Guiana
Tropical Forests of the Guiana Shield, Ancient Forests in a Modern World
Vegetation Map of Guyana, 1:1,000,000°
Amazon rainforests green-up with sunlight in dry season
Geophysical Research Letters
Manual técnico da vegetação brasileira
Mapa de Vegetação
Tree species diversity and forest structure in relation to microtopography in a tropical freshwater swamp forest in French Guiana
Plant Ecology
Cited by (57)
Predicting the loss of forests, carbon stocks and biodiversity driven by a neotropical ‘gold rush’
2023, Biological ConservationEcology, evolution, and epidemiology of zoonotic and vector-borne infectious diseases in French Guiana: Transdisciplinarity does matter to tackle new emerging threats
2021, Infection, Genetics and EvolutionCitation Excerpt :These images have been used to map land cover and land use (LCLU), all involving human intervention at some point in the classification process, particularly due to the significant cloudiness in the region limiting the applicability of fully automatic methods. In addition to these LCLU maps, several already available remote sensing-derived products were used: a digital elevation model and geomorphological landscapes from the Shuttle Radar Topography Mission (SRTM) (Guitet et al., 2013); land use from the VEGETATION sensor onboard the SPOT-4 satellite (1-km spatial resolution) (Gond et al., 2011); and the human footprint (de Thoisy et al., 2010). As examples, landscape characterization has made it possible to objectively design the sampling for the capture of Anopheles, with the aim of inventorying species (Roux et al., 2013).
From inselberg to inselberg: Floristic patterns across scales in French Guiana (South America)
2017, Flora: Morphology, Distribution, Functional Ecology of PlantsCitation Excerpt :A noticeable difference in floristic composition (beta diversity) was reported between the forest located in the northern part and the southern part of French Guiana, in relation to geomorphology (Sabatier and Prévost, 1990), and a general pattern may be described at a regional scale (Gond et al., 2011; Guitet et al., 2015). At present, southern forests, with their lower level of annual rainfall, are characterized by their almost open aspect (discontinuous canopy), especially in the southern peneplain where numerous local clearings are found, while northern forests show a continuous canopy typical for rain forests (Gond et al., 2011). Elevation may also yield contrasting effects on floristic composition.
- 1
Current address: Joint Research Centre of the European Commission, Global Environement Monitoring Unit (GEM), Via E. Fermi, 2749 I-21027 Ispra (Va) Italy.