|
" Utilization of satellite and close range digital data for mapping soils and soil properties in arid rangeland and agricultural rainfed and irrigated lands, Morocco "
M. Naimi
P. C. Robert
| Document Type
|
:
|
Latin Dissertation
|
| Language of Document
|
:
|
English
|
| Record Number
|
:
|
1113296
|
| Doc. No
|
:
|
TLpq304471567
|
| Main Entry
|
:
|
M. Naimi
|
|
|
:
|
P. C. Robert
|
| Title & Author
|
:
|
Utilization of satellite and close range digital data for mapping soils and soil properties in arid rangeland and agricultural rainfed and irrigated lands, Morocco\ M. NaimiP. C. Robert
|
| College
|
:
|
University of Minnesota
|
| Date
|
:
|
1998
|
| student score
|
:
|
1998
|
| Degree
|
:
|
Ph.D.
|
| Page No
|
:
|
274
|
| Abstract
|
:
|
SPOT multispectral, spectroradiometric and ERS-1 radar data were used for mapping soils at Aarid and Sais sites in Morocco. Spectral curves at the Aarid site showed differences between soils and a significant amount of variability within soils. The reflectance variation was explained by the crusting, stoniness, and carbonate content of the soil surface. Although these factors explained some reflectance variation, intrinsic factors were used in regression models to predict color hue (R2 = 0.34), and carbonates (0.66). Spectral reflectance differences at the Sais site were insignificant to discriminate soil types. Similar spectral behavior and low amplitude between soil curves were striking. Besides, soil intrinsic factors seemed meaningful in explaining reflectance variability. But, only sand and clay, carbonate, CEC, iron oxides and organic matter could be predicted. Extrinsic soil factors, such as tillage, crop residues and stoniness determined the reflectance. Use of SPOT imagery in mapping Aarid soils allowed us to conclude that: more than one soil transect was needed; primary SPOT data were as good as enhanced bands in mapping soils and land cover; mapping soils in locations away from the training site locations was difficult. Identification of soils using vegetation was very reliable for Stipa-covered soils. Although combined dual SPOT data was more suitable to depict linear features and geometric polygons, merged data improved significantly visual quality and image spatial detail for refining soil and land cover boundaries. Despite spectral differences between soils at the Sais site, soil mapping using multispectral data was very difficult because soil natural boundaries interfere with artificial borders of parcels. The integration of digital elevation model helped overcome limitations of SPOT digital data. Relationships between ERS-1 and soil surface moisture and roughness were studied. Simple regressions between and moisture was significant when wetness was expressed as AWC fraction for sugarcane, bare soil, and bare soil plus sunflower (0.60*, 0.88**, 0.85**). When was related to roughness parameters, R2 was 0.51* for sugarcane fields, and 0.61* when 1/S2 was used (0.61*). Multiple regressions R2 was 0.49* for sugarcane and 0.73* for bare soil. The use of transformed variables improved the relationship (0.74*, 0.55*).
|
| Subject
|
:
|
Agricultural lands
|
|
|
:
|
Applied sciences
|
|
|
:
|
Arid rangeland
|
|
|
:
|
Biological sciences
|
|
|
:
|
Close range digital data
|
|
|
:
|
Irrigated
|
| |