Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Research Project #440729

Research Project: CEAP Wetlands Mid-Atlantic Coastal Plain Regional Study

Location: Hydrology and Remote Sensing Laboratory

Project Number: 8042-13610-030-044-I
Project Type: Interagency Reimbursable Agreement

Start Date: Oct 1, 2020
End Date: Sep 30, 2024

Wetlands in agricultural landscapes are an important source of ecosystem services and represent a substantial portion carbon storage in agroecosystems. This research will address these aspects. 1. Characterize ecosystem service provisioning and mapping quality for soil organic carbon in natural and restored wetlands. 2. Impacts of riparian buffers and wetland restoration on pollination ecosystem service provisioning. 3. Mapping ditch drainage network based on deep learning application to lidar data. 4. Wetland classification using Sentinel-1 C band SAR based on a CNN-based denoising module. 5. Case study applications of a remote sensing-based wetland functional assessment.

Objective 1: We will explore the usage of common derived elevational terrain attributes as predictor variables for SOC mapping for agricultural landscapes dominated by prior converted cropland. With the development of light detection and ranging (lidar) technology, the availability of high-resolution (e.g., < 3m) lidar data is increasing, which allows the characterization of subtle changes in local terrain for detecting the spatial variability of soil organic carbon (SOC). Objective 2: The objective of this study is to assess the impacts of riparian buffers and wetland restorations on pollinator services in the Chesapeake Bay Watershed. Objective 3. This study will apply a state-of-the-art semantic segmentation deep learning method to map the ditch network over the Upper Choptank River watershed using high-resolution lidar-derived topographic data. Objective 4. The aim of this study is to evaluate an innovative Convolutional Neural Network (CNN)-based time series SAR denoising module, based on multiple dates Sentinel-1 C band SAR in Delaware in 2017, in conjunction with traditional pixel-based (i.e. random forest) and the recent deep learning based (i.e., U-net) classification approaches. Objective 5. Long-term objective of case studies will be a finalized protocol to determine the efficacy of the assessment in the creation of ecosystem service inventories and conservation applications. At the watershed scale, the assessment will be applied to NWI wetlands across a selected watershed to inventory wetland functions and associated ecosystem services.