Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #317929
Title: Mapping tillage intensity by integrating multiple remote sensing data

Author
item Gao, Feng
item Daughtry, Craig
item Stern, Alan
item QUEMADA, MIGUEL - Polytechnic University Of Madrid

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 6/20/2015
Publication Date: 9/14/2015
Citation: Gao, F.N., Daughtry, C.S., Stern, A.J., Quemada, M. 2015. Mapping tillage intensity by integrating multiple remote sensing data [abstract]. 20th International Soil Tillage Research Organization (ISTRO) Conference, Nanjing, China, September 14-18, 2015. 2015 CDROM.

Interpretive Summary:

Technical Abstract: Tillage practices play an important role in the sustainable agriculture system. Conservative tillage practice can help to reduce soil erosion, increase soil fertility and improve water quality. Tillage practices could be applied at different times with different intensity depending on the local weather condition and farmer’s assessment. Remote sensing approach has been applied to identify tillage intensity at field scales by mapping crop residues left on the soil surface shortly after planting. Medium spatial resolution remote sensing data such as Landsat (30m) provides residue information at field scale from the two shortwave infrared bands. However, remote sensing data may not be available shortly after the planting date due to long revisit cycle (16 days for Landsat) and cloud contamination etc. Efforts to assess tillage intensity using a single date of Landsat data have been only moderately successful. A multi-temporal approach has been demonstrated more robust in different cropping regions of the United States. In order to build high spatial and temporal information that are required for mapping tillage intensity, we introduce data fusion approach by integrating multiple remote sensing sources. The Moderate Resolution Imaging Spectroradiometer (MODIS) (daily revisit, 500m pixel resolution) are fused with Landsat data to generate the daily 30m pixel resolution reflectance using the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM). We test the hypothesis that the fused Landsat and MODIS data provides more accurate information on mapping crop residue and tillage intensity than using individual sensor imagery alone. The study area covers the South Fork and Walnut Creek (USDA Conservation Effects Assessment Project, CEAP) watershed in central Iowa USA. Over the study sites, crop residue cover was measured in more than 50 fields per year using the line point transect method and estimated visually in more than 200 fields/year using roadside survey methods. We used partial measurements to train remote sensing data and then evaluated results using the remaining samples. Figure below demonstrates the procedure for mapping tillage intensity using data fusion approach. Results based on single Landsat, dense time-series MODIS and the fused Landsat-MODIS are compared and analyzed using the 2011 and 2013 field measurements. We will discuss the potential and limitation of the data fusion approach for the tillage intensity mapping in this presentation.