Integrating UAV optical imagery and LiDAR data for assessing the spatial relationship between mangrove and inundation across a subtropical estuarine wetland
Graphical abstract
Introduction
Mangrove forests, one of the most important coastal blue carbon ecosystems, are distributed within limited intertidal zone typically between mean seal level (MSL) to highest spring tide in coastal tropical and subtropical climates in the world (Alongi, 2009, Giri et al., 2010). Mangrove forests provide critical ecosystem services such as protection from erosions, carbon sequestration, maintaining biodiversity (Costanza et al., 1997, Donato et al., 2011, Alongi, 2014). However, mangrove forests are now threatened globally with significant loss in area and severe degradation in habitats mainly due to human activities including land conversion to aquaculture, agriculture, urban development (Valiela et al., 2001, Giri et al., 2010). Future climate change, sea level rise and more intensive human activities will much likely make it worse for mangrove forests. Therefore, sound conservation and management practices of mangrove forests are highly needed. To achieve this, the first step is to enhance our limited understanding of mangrove ecology including the key factors influencing the distribution and growth of mangrove species.
The distribution of mangrove species in the intertidal zone is subjected to multiple environmental gradients (inundation, salinity, soil texture, nutrients, etc.) and the spatial interactions of these gradients are complex and site-dependent (Krauss et al. 2008). Among these proposed environmental drivers, tidal inundation regime (i.e., flooding frequency and duration) has been widely reported as a key driver governing mangrove species distribution (Chen et al., 2013, Crase et al., 2013, Spier et al., 2016, Leong et al., 2018), since tidal activities are relatively easy to be documented and are closely linked to other drivers including salinity, soil texture, and redox potential. Indeed, the link between tidal inundation and species distribution has been widely discussed in mangrove ecology for nearly a century, asserting that coastal mangrove forests exhibit zonation of species along the elevation gradient (Watson, 1928, Chapman, 1976, Snedaker, 1982). The relationships between tidal inundation regime and mangrove species distribution are appealing to mangrove ecologists in three aspects: (1) to understand underlying mechanisms of mangrove forests adapting to flooding intertidal zone (Naidoo, 1983, Chen and Wang, 2016), (2) to support mangrove restoration practices by indicating appropriate elevation ranges for different mangrove species (Kodikara et al., 2017, Oh et al., 2017), and (3) to assess the vulnerability and resilience of mangrove forests to future sea level rise (Nitto et al., 2014, Lovelock et al., 2015). There are already a number of laboratory and field studies exploring the inundation-mangrove relationships, but the influence of inundation regime on mangrove forests is still poorly understood and quantitative mangrove-inundation relationships have not been well documented (Krauss et al. 2006).
To analyze mangrove-inundation relationships in a quantitative manner, it is essential to accurately measure the spatial variability of tidal inundation regime; however, this is challenging in the intertidal zone vegetated with mangrove forests which have a flat landscape but complex microtopography (e.g., small creeks and basins). The characterization of microtopography is important since the distribution and growth of mangrove species are affected by small differences in elevation (Oh et al. 2017). Tidal gauge measurements can capture the temporal dynamics of tidal inundation regime at single-point scales, but they are not able to characterize the spatial variability of tidal inundation regime across the intertidal zone. In many of field studies exploring the inundation-mangrove relationships, single-point measurements of tidal water level and ground-based spatial surveying of surface elevation with high-accuracy instruments (such as real-time kinematic GPS) are often combined to estimate the spatial variability of inundation regime (Ellison et al., 2000, Crase et al., 2013, Leong et al., 2018), assuming that intertidal zone of equal surface elevation have the same inundation regime. The ground-based surface elevation surveying is time-consuming and labor-intensive due to the inaccessibility of many mangrove forests. It might be suitable for those studies only focusing on several mangrove transects (e.g., Ellison et al. 2000); however, it is much challenging to apply this method to characterize the spatial variability of surface elevation across the intertidal zone with complex microtopography particularly under dense mangrove canopies.
Many remote sensing data, such as SRTM, ICESat/GLAS, TanDEM-X, and ALOS PALSAR, from spaceborne interferometric synthetic aperture radar (InSAR) and light detection and ranging (LiDAR) instruments have been used to characterize large-scale spatial variability of terrain and vegetation information in coastal wetlands (Cornforth et al., 2013, Fatoyinbo and Simard, 2013, Kovacs et al., 2013, Lee and Fatoyinbo, 2017). As examples, Fatoyinbo and Simard (2013) estimated canopy heights of all mangroves in Africa at continental scale from SRTM and ICESat/GLAS data, and Lee and Fatoyinbo (2017) developed inversion approaches using TanDEM-X data to estimate mangrove canopy heights showing the possibility of global-scale applications. These remote sensing data from spaceborne instruments have shown great potentials for generating large-scale canopy height model (CHM) of mangrove forests, but they have relatively low spatial resolutions (tens of meters) and height accuracy (error in meters) (Lucas et al. 2017), which restricts them from local-scale studies. At local scales, stereo photogrammetry using very high resolution (VHR) images from unmanned airborne vehicle (UAV) and satellites has been used to generate CHM of mangrove forests (e.g., Lucas et al. 2002), but this method still suffers from low accuracy (error in meters) in estimated mangrove canopy heights.
With the developments of UAV and LiDAR technologies and the reduce in the expense of applying these technologies, UAV LiDAR remote sensing has been increasingly used to characterize coastal terrains and vegetation structures (Morris et al., 2005, Lagomasino et al., 2016, Fatoyinbo et al., 2017, Wannasiri et al., 2013). On the one hand, the ability of UAV LiDAR to measure surface elevation with high accuracy even under dense forest canopies makes it ideal to detect the spatial variability of microtopography over the intertidal zone (Crase et al. 2013), which is then combined with tidal water level measurements to map the spatial distribution of inundation regime. On the other hand, the ability to acquire forest canopy heights at a centimeter accuracy level from LiDAR-derived CHM (Feliciano et al. 2017) provides an ideal opportunity to examine the influences of inundation regime on the growth of mangrove species at the landscape scale. Although airborne LiDAR has been often proposed as the most accurate way for forest structure measurements in various forest types (Zhao et al., 2012, Zolkos et al., 2013, Duncanson et al., 2015), only a few airborne LiDAR applications have been done so far for mangrove forests and the practicability of these applications has not been well verified.
In this study, we integrate UAV camera and LiDAR data for exploring mangrove-inundation spatial patterns across a subtropical estuarine wetland. The main goal of this study is to quantitatively evaluate the role of inundation regime in governing the spatial distribution and growth of different mangrove species at the landscape scale. Our specific objectives are to examine (1) how different mangrove species occur along the gradients of elevation and inundation, and (2) how the mangrove-inundation spatial patterns differ among species.
Section snippets
Study area
The study area is a subtropical intertidal wetland with an area of ∼2.6 km2, located to the south of Zhangjiang estuary in the southeast coast of China (Yunxiao, Fujian, China; Fig. 1a). This region has a mangrove area of ∼0.5 km2 mainly including Kandelia obovate (K. obovata), Avicennia marina (A. marina), and Aegiceras corniculatum (A. corniculatum), as well as invasive Spartina alterniflora (S. alterniflora). The three main mangrove species have distinct color characteristics in early summer
Spatial distributions of mangrove species and inundation regime
Based on high-resolution spatial data of UAV camera and LiDAR CHM, the orthomosaic image was classified into different land covers including three main mangrove species, S. alterniflora, mudflat, and tidal creeks (Fig. 2a). The land cover map had an overall classification accuracy of 78.6% across different land covers and mean producer’s and user’s accuracy of 83.6% and 86.9% respectively for mangrove species (Table 1). Almost all mangrove forests over Zhangjiang estuarine wetland were situated
Discussion
Our results confirm that the inundation regime plays a critical role in governing the spatial distributions of mangrove species. In accord with general understanding of elevation distributions for mangrove forests (Macnae, 1967, Lear and Turner, 1977, Alongi, 2009), almost all mangrove forests in Zhangjiang estuarine wetland were distributed over the intertidal zone with relative elevation between MSL and HHW (Fig. 5). As a consequence, mangrove forests were on average inundated for < 12 h/day
Conclusions
Spatially explicit quantitative analyses were conducted in this study to examine the mangrove-inundation spatial patterns, including the influences of inundation on spatial distribution and canopy height of different mangrove species, over the intertidal zone in a subtropical wetland, i.e., Zhangjiang estuarine wetland, Fujian, China. The mapping involved in our analyses, including mangrove species (K. obovate, A. marina and A. coniculatum), canopy height, inundation period, and relative
Acknowledgements
We thank B. Huang and Y. Zhang for their suggestions on designing this study, and thank L. Meng, P. Lin, K. Chen, and C. Zheng for their help in the field work. We also would like to thank the Zhangjiang Estuary Mangrove National Nature Reserve for their long-term support of our ecological research programs. This study was supported by the National Natural Science Foundation of China (31600368), the Natural Science Foundation of Fujian Province, China (2017J01069), the Fundamental Research
References (57)
Mangrove zonation: an examination of data from seventeen riverine estuaries in tropical Australia
Annals Botany
(1996)- et al.
A mangrove forest map of China in 2015: analysis of time series landsat 7/8 and sentinel-1A imagery in Google earth engine cloud computing platform
ISPRS J. Photogrammetry Remote Sens.
(2017) - et al.
Significant niche overlap between native and exotic Sonneratia mangrove species along a continuum of varying inundation periods
Estuar. Coast. Shelf Sci.
(2013) - et al.
An object-based method for mapping and change analysis in mangrove ecosystems
ISPRS J. Photogramm. Remote Sens.
(2008) - et al.
The importance of spatial detail: Assessing the utility of individual crown information and scaling approaches for lidar-based biomass density estimation
Remote Sens. Environ.
(2015) - et al.
UAV photogrammetry for topographic monitoring of coastal areas
ISPRS J. Photogramm. Remote Sens.
(2015) - et al.
The mangrove forest dynamics model mesoFON
Ecol. Model.
(2014) - et al.
Applications of ALOS PALSAR for monitoring biophysical parameters of a degraded black mangrove (Avicennia germinans) forest
ISPRS J. Photogramm. Remote Sens.
(2013) - et al.
Environmental drivers in mangrove establishment and early development: a review
Aquat. Bot.
(2008) - et al.
High-resolution pattern of mangrove species distribution is controlled by surface elevation
Estuar. Coast. Shelf Sci.
(2018)
Integration of classification methods for improvement of land-cover map accuracy
ISPRS J. Photogramm. Remote Sens.
The role of surface elevation in the rehabilitation of abandoned aquaculture ponds to mangrove forests, Sulawesi, Indonesia
Ecol. Eng.
Flood regime as a driver of the distribution of mangrove and salt marsh species in a subtropical estuary
J. Mar. Syst.
Allometric equation choice impacts LiDAR-based forest biomass estimates: A case study from the Sierra National Forest, CA
Agric. For. Meteorol.
A meta-analysis of terrestrial aboveground biomass estimation using lidar remote sensing
Remote Sens. Environ.
Carbon cycling and storage in mangrove forests
Annual Rev. Marine Sci.
The Energetics of Mangrove Forests
Object-based mangrove species classification using unmanned aerial vehicle hyperspectral images and digital surface models
Remote Sensing
Mangrove vegetation
J Cramer
Ecophysiological responses of viviparous mangrove rhizophora stylosa seedlings to simulated sea-level rise
J. Coastal Res.
Photosynthetic and physiological responses of Kandelia candel L. Druce seedlings to duration of tidal immersion in artificial seawater
Environ. Exp. Bot.
Influence of water logging time on the growth of Kandelia candel seedlings
Acta Oceanologica Sinica
Advanced land observing satellite phased array type L-band SAR (ALOS PALSAR) to inform the conservation of mangroves: Sundarbans as a case study
Remote Sensing
The value of the world’s ecosystem services and natural capital
Nature
Hydroperiod is the main driver of the spatial pattern of dominance in mangrove communities
Glob. Ecol. Biogeogr.
Mangroves among the most carbon-rich forests in the tropics
Nat. Geosci.
Testing patterns of zonation in mangroves: scale dependence and environmental correlates in the Sundarbans of Bangladesh
J. Ecol.
Cited by (71)
Strong diurnal variability of carbon dioxide flux over algae-shellfish aquaculture ponds revealed by eddy covariance measurements
2023, Agriculture, Ecosystems and EnvironmentMethodology for fine-scale seamless elevation model construction using unoccupied aerial and surface vehicles
2023, Estuarine, Coastal and Shelf Science