Water Resources Research Act Program

Details for Project ID 2004MD57B

Leaf-scale hyperspectral reflectance models for determining the nitrogen status of freshwater wetlands

Institute: Maryland
Year Established: 2004 Start Date: 2004-03-01 End Date: 2005-03-01
Total Federal Funds: $29,957 Total Non-Federal Funds: $61,093

Principal Investigators: David Tilley, Andrew Baldwin

Project Summary: Problem: The eutrophication of open-water aquatic systems, such as the Chesapeake Bay, from non-point source (NPS) pollution is increasingly viewed as a global threat to ecosystem health. Tidal freshwater marshes (TFM), due to their position in the landscape between land that generates NPS and open-water aquatic systems that receive it, serve as natural nutrient filters that intercept NPS, reducing its potential impact. Detection of nutrient additions to TFMs could serve as a sentinel of NPS pollution. Development of a remote sensing technique that can quickly scan TFMs for the effect of NPS could help environmental managers locate nutrient hot spots. We are developing a boat-borne hyperspectral radiometer as a tool for remotely sensing the nutrient status of TFMs. Our previous research demonstrated the proof-of-concept of using leaf-scale hyperspectral reflectance models for assessing marsh water column nitrogen (N) status. A recently funded award from MD Sea Grant is being used to design, build and prove the concept that a boat-radiometer system can collect canopy reflectance in TFMs. Funding from MWRRC will be used for the next phase of reflectance model development, which is to examine in variation of canopy reflectance among species and across nitrogen gradients. We have also requested twoyears of funding from NOAAs CICEET program that would fund pilot-scale testing of the boat-radiometer system in a variety of TFMs, which are influenced by various environmental factors, to improve data collection and interpretation methods. Objectives: Our objective for this project is to test the applicability of previously developed leaf-scale reflectance models over a broader nitrogen range and with other wetland species. The specific objective is to experimentally examine how narrow spectral band reflectance indices (e.g., PRI, RE, R678/R493) vary between three common species of wetland plants across a nitrogen gradient. This objective is an integral component of our larger goal of developing a boat-borne radiometer system for identifying nonpoint source pollution in wetlands. Methodology: Three wetland plant species (Typha latifolia, Phragmites australis, and Acorus calamus), grown in containers in a greenhouse, will be treated with prescribed quantities of N to achieve a desired pore water N level. Three replicates of each species will be arranged in a completely randomized design. An ASD Handheld SpectroRadiometer (Analytical Spectral Devices, Boulder, CO), positioned directly above the potted plant, in direct sunlight on cloud-free days will measure plant reflectance. We will analyze experimental data as a two-way repeated measures Analysis of Variance (ANOVA). Two separate ANOVA models will be used. One model will treat species and N level as categorical variables so that treatment means can be calculated and means separation procedures applied. The second model will treat N level as a continuous variable so that regression equations describing the relationship between N level reflectance can be developed for each species. Rationale: Advancing the capability of wetland remote sensing to quantify the nitrogen status of wetlands can (1) provide a tool for the large scale monitoring of water quality in difficult-to-access wetlands, (2) offer a rapid screening method for identifying nitrogen hot-spots in a watershed, and (3) enable near real-time monitoring in areas suspected of producing significant quantities of non-point source (NPS) pollution.