An evaluation of DEMs derived from LiDAR and photogrammetry for wetland mapping

Abstract

The Ontario Ministry of Natural Resources (OMNR) and Ducks Unlimited Canada (DUC) have been engaged in developing an efficient and accurate methodology for inventorying wetlands. Their progress in this area has demonstrated that Digital Elevation Models (DEMs) are crucial input for wetland identification and boundary delineation. The provincial DEM, however, has known precision limitations in areas of minimal topographic relief that cause considerable mapping error. This study explored whether wetland mapping derived from bare-earth light detection and ranging (LiDAR) data would overcome the limitations of the provincial DEM. An automated wetland mapping approach was applied to the 2 elevation datasets and the results were compared using 2 methods of validation. One hundred aerial-photo-interpreted sample plots were used to quantitatively measure the ability of each source to separate upland from wetland. An overlay of wetland maps created from the 2 DEM sources was then qualitatively assessed to further clarify the magnitude of discrepancy between the 2 mapping sources. The study concluded that LiDAR showed a significant improvement at p = 0.05 over the provincial DEM for mapping wetlands, improving overall mapping accuracy from 76% to 84%. However, an overlay analysis and qualitative assessment showed the magnitude of this reported improvement is greater than was quantified by the accuracy assessment and that an assessment scheme with different sample units may further elucidate this discrepancy.Key words: LiDAR, DEM, wetland, mapping RÉSUMÉLe Ministère des Richesses naturelles de l'Ontario (MRNO) et Canards illimités Canada (CIC) ont entrepris le développe-ment d'une méthodologie efficace et précise d'inventaire des milieux humides. Les progrès enregistrés dans ce domaine indiquent que les modèles de numérisation de l'altitude (MNA) constituent un élément clé de l'identification des milieux humides et de leur cartographie. Le MNA provincial, cependant, a révélé des limites dans la précision des zones à relief peu accidenté qui ont entraîné d'importantes erreurs cartographiques. Cette étude a cherché à savoir si la cartographie des milieux humides dérivée des données de détection de la lumière et de calcul de la distance (LiDAR) sur zone dénudée pouvait permettre de pallier les limites du MNA provincial. Une approche automatisée de cartographie des milieux humides a été utilisée pour deux ensembles de données altimétriques et les résultats ont été comparés au moyen de deux méthodes de validation. Cent parcelles-échantillons interprétées à partir de photos aériennes ont été utilisées pour mesurer quantativement la capacité de chaque source à séparer les zones sèches des milieux humides. Une superposition des cartes de milieux humides produites à partir de deux sources de MNA a été par la suite évaluée qualitativement pour identifier plus en détail l'importance des différences entre les deux sources de cartographie. L' étude permet de conclure que le LiDAR a démontré une amélioration significative, p =. ...

Google Earth Engine, Open-Access Satellite Data, and Machine Learning in Support of Large-Area Probabilistic Wetland Mapping

“…Comparison with other models, as well as our reference data set, indicate that overall patterns of wetland distribution are well-captured within our study area. The TWI input variable was of greatest importance to our T model (Figure 3), reflecting the importance of local topographic and hydrographic conditions in wetland occurrence: a relationship also demonstrated by other topographically-based wet areas mapping efforts such as those described by [82,83]. Indeed, the TWI showed the greatest influence of all input variables in each of the four probability models, regardless of what additional input variables were included (Figure 3).…”

The missing wetlands: using local ecological knowledge to find cryptic ecosystems

“…For example, 36 of the 44 LEK wetlands in our study had no distinguishable DEM signature for DEMs derived from contour data. Several studies have demonstrated the potential of LiDAR in detecting wetlands (e.g., Hogg and Holland 2008;Julian et al 2009;Lang and McCarty 2009;Maxa and Bolstad 2009). These higher resolution digital resources are available for some areas and, at least in low-relief landscapes including Atlantic coastal plain pine forests, hold much promise for detecting isolated wetlands (e.g., P. Leonard, ''unpublished data'').…”

A Geospatial Approach for Identifying and Exploring Potential Natural Water Storage Sites

“…Aircraft-mounted LiDAR sensors measure groundcover characteristics by determining the travel time of laser pulses (hundreds of thousands per second) to objects on the ground and back to the airborne sensor. The current generation of high resolution DEMs and digital surface models make LiDAR valuable for forestry and vegetation analyses, while the ability to filter and remove the influence of vegetation canopies make LiDAR especially useful in wetland research applications [56][57][58]. Use of LiDAR data in wetland studies eliminates the need for in situ laser transects and the high costs associated with regular field visits [43,[58][59][60][61].…”

“…The current generation of high resolution DEMs and digital surface models make LiDAR valuable for forestry and vegetation analyses, while the ability to filter and remove the influence of vegetation canopies make LiDAR especially useful in wetland research applications [56][57][58]. Use of LiDAR data in wetland studies eliminates the need for in situ laser transects and the high costs associated with regular field visits [43,[58][59][60][61]. In Montana, LiDAR elevation data have been particularly effective when employed in flood inundation research, human stream impact studies, and channel migration zone mapping [62][63][64][65][66].…”