Objective
Extract quantitative biophysical information about vegetation and soils by exploiting the spectral, spatial, temporal, and bidirectional domains of remotely sensed data.

This objective is directly responsive to the Soil Resource Management Program (NP 202) and Water Quality and Management Program (NP 201).

CROP NUTRIENTS AND NUTRIENT BUDGETS
-- Leaf to Field Scale --

Nitrogen (N) is an essential element for plant growth and is frequently the major limiting nutrient. Farmers must balance the competing goals of supplying adequate N for their crops against minimizing N losses to the environment.

Currently, soil sampling and leaf chlorophyll meters (SPAD) are used to predict crop nutrient requirements. Site-specific nutrient management depends on identifying the spatial variation of soil nutrients, particularly N, and applying appropriate nutrients to optimize crop growth and yield. Active (fluorescence) and passive (reflectance) remote sensing techniques can provide information on leaf chlorophyll concentration that is related to plant growth.

Fluorescence Methods:Steady state laser-induced fluorescence is related to the concentration and photosynthetic activity of the plant pigments. Excitation with UV light (337-400nm) of corn plants fertilized at different nitrogen levels produced differences in the fluorescence at 440, 525, 685, and F740nm. A 440/685 ratio can produce significant separations. What appears to be most encouraging is that the data trends tended to even separate the optimal recommended rate of fertilization for over fertilized rates as well as under fertilized rates.

(Left Figure) Multispectral laser induced fluorescence imaging system (LIFIS). (Right Figure) Nitrogen fertilization produces differences in leaf and canopy fluorescence.

Quantifying crop residue cover is important for evaluating the effectiveness of conservation tillage practices. Current methods of measuring residue cover are inadequate for characterizing the spatial variability of residue over large fields. The reflectance of crop residues at a particular visible or near infrared (400-1100 nm) wavelength may be higher or lower than the reflectance of the soil. This makes discrimination between crop residues and soils difficult or nearly impossible using reflectance techniques alone in the visible and near infrared wavelengths.

Click on image to enlarge

Reflectance Methods

As the scale expands from leaf to field, the signal from remotely sensed measurements is affected by attributes of different components such as plant architecture, soil color, and soil texture. Other independent factors such as weather, air pollution, and levels of atmospheric carbon dioxide play an increasing complex role because these factors affect photosynthesis and growth.

Variations in soil reflectance and leaf area index often confound the assessment of leaf N by remote sensing techniques. Thus, attempts to assess plant nutrient status based on canopy reflectance are often confounded by the variability in background reflectance and/or leaf area index. Spectral indices are needed that are sensitive to leaf chlorophyll concentration and that minimize variations in canopy reflectance associated with changes in background reflectance and LAI.

-- Field to Watershed Scale --

Agriculture is a potentially strong source of nutrients in surface and ground water. Site-specific production systems are advocated as superior alternatives to conventional farming practices because of the potential to minimize detrimental impacts on the environment.

The primary experimental site is the 2500-hectare Beltsville Agricultural Research Center (BARC) in Maryland. Within BARC, several intensive and extensive test sites have been identified and extensive databases are being developed.

The intensive test sites are part of a multi-disciplinary project entitled Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3). The primary focus of OPE3 is to develop farming strategies that conserve natural resources while maintaining or increasing long-term farm profitability.

The infrastructure of OPE3 provides for meaningful comparisons of several agricultural production systems because the research site is large enough to capture the spatial variability of crop and soil parameters, yet not so large that the fields themselves are in different climatic or geologic settings. The site has four hydrologically bounded watersheds, about 4 ha each, which feed a wooded riparian wetland and first-order stream.

Three watersheds have similar surface and subsurface soil and water flow characteristics and will be used to evaluate three crop production systems:

1. uniform application of agricultural chemicals
2. site-specific application of agricultural chemicals
3. uniform applications of dairy manure

The remaining watershed has similar soil characteristics, but differs significantly in subsurface water flow patterns. This provides a unique opportunity to evaluate the same crop production system on watersheds with contrasting subsurface flow patterns.

The extensive test sites consist of the numerous grain and forage crop production fields, as well as forests and wetlands, that surround the intensive test sites at BARC. Additional sites with known management histories at other ARS and/or collaborator locations will be included as appropriate.

CROP RESIDUE COVER

Quantifying crop residue cover is important for evaluating the effectiveness of conservation tillage practices. Current methods of measuring residue cover are inadequate for characterizing the spatial variability of residue over large fields. The reflectance of crop residues at a particular visible or near infrared (400-1100 nm) wavelength may be higher or lower than the reflectance of the soil. This makes discrimination between crop residues and soils difficult or nearly impossible using reflectance techniques alone in the visible and near infrared wavelengths.

Click on image to enlarge
(Left Figure) Reflectance spectra of wheat residues and Othello soil at a range of relative water contents from oven dry (RWC = 0) to water saturated (RWC = 1). Two very broad absorptions centered near 1600 nm and 2100 appear in all compounds possessing alcoholic -OH groups, such as, sugars, starch, and cellulose. These absorption features are present in crop residues, but not in soils. (Right Figure) Cellulose absorption index (CAI) of corn, soybean, and wheat residues and soils plotted as function of relative water content (RWC). The dashed line is at CAI = -1.0. All soils are located below the line and all crop residues are located above the line.

The wide range of CAI values makes quantification of crop residue cover feasible. This reflectance technique appears promising for field and regional-scale surveys.

WEED AND NARCOTIC CROP DETECTION

The fundamental hypothesis is that weed or marijuana plants have something unique about their spectral, spatial, temporal, or bidirectional reflectance characteristics. We are exploring these domains via laboratory, field, and airborne measurements.

Knowledge of the differences in leaf reflectances is a useful starting point when looking for features to discriminate between species using spectral remote sensing. The amount of variability and the significance of the variability within and among species is still not understood.

Click on image to enlarge
Differences in the reflectance spectra of tree leaves from the reflectance spectrum of marijuana leaves. The line at zero (0) indicates no change from the reflectance of marijuana.

Color photographs of a marijuana garden.

CROP RESPONSE ZONES

Soil characteristics that determine field variability can be mapped by assessing crop productivity via remote sensing and crop growth simulation models. Contiguous locations within a field that co-vary because of similar responses to environmental conditions such as soil/precipitation interactions can be aggregated to form response zones. A crop growth model can be used to link productivity measures from remote sensing imagery to soil water holding capacity for different precipitation patterns. Once delineated, response zones can form the basis for creating management zones. The study seeks to use readily available data.

   
Although drought conditions were more severe in 1999 than in 1998, yield maps from both years, 1998 (Left Figure) and 1999 (Right Figure), showed a marked degree of similarity in the spatial pattern of the higher producing areas.

   
(Left Figure) Potential groundwater flow pathways determined from surface topography and depth of subsurface restricting layers at watersheds A and B. (Right Figure) High yielding areas (dashed lines) were closely correlated with areas over subsurface flow pathways at watersheds A and B.

Crop Nutrients

• Robust methods to relate leaf optics to leaf nitrogen and crop yield.
• Fluorescent systems to assess plant stress and photosynthesis at the field level.
• Robust reflectance methods to normalize for differences in soil brightness and LAI and extract information on leaf optics.
• Reliable methods to scale nitrogen budgets from leaf to field and landscape levels.

Crop Residue Cover

• Low cost, reliable instrument to replace line-transect.
• Methods to assess effectiveness of conservation tillage over large areas.
• Methods to scale measurements of soil organic carbon to landscape level.

Weed Detection

• Accurate 3-D canopy reflectance models.
• BRDF characterization for different species in heterogeneous situations.
• Airborne systems that acquire images at very high spatial resolution (<0.5m) or off-nadir view angles.

Crop Response Zones

• LAI retrieval often fails at high spatial resolution.
• Weather frequently hinders airborne data collection on the east coast.
• Decoupling of water available in the root zone may indicate subsurface soil structure that is affecting movement of water and nutrients, and hence production.

• Alternative sensor systems, such as LIDAR and thermal infrared at high spatial resolution.
• Optimization of inputs including antecedent Soil Water Holding Capacity (SWHC) characterization for crop growth model initialization.
• Quantification of resolution limits for detectability and signature separability for classifications.
• Develop sensor recommendations for crop residue monitor.
• Develop algorithms for regional surveys of crop residue and crop response zones.

National Program 207, Integrated Agricultural Systems

Collaborators:

• Hydrology Laboratory (HL), BARC
• Environmental Chemistry Laboratory (ECL), BARC
• National Soil Tilth Laboratory, Ames, IA
• NASA, Goddard Space Flight Center, Greenbelt, MD
• University of Maryland, College Park, MD
• 3DImaging, Inc., Easton, MD
• National Guard Bureau