Narrow your search

Library

KBR (1)

KU Leuven (1)


Resource type

dissertation (1)


Language

English (1)


Year
From To Submit

2007 (1)

Listing 1 - 1 of 1
Sort by

Dissertation
Recyclable nano-gold as chemoselective catalyst for liquid phase oxidations and hydrogenations
Authors: ---
ISBN: 9789088260384 Year: 2007 Volume: 771 Publisher: Leuven Katholieke Universiteit Leuven

Loading...
Export citation

Choose an application

Bookmark

Abstract

Within-field variability of the plant-available nutrients often results in different fertiliser requirements across a field. Variable rate (VR) fertilisation is one of the advanced techniques in precision agriculture. The VR strategy for P (VRP) is based on standard soil test for P in sampling grid of one hectare for large agricultural fields. The study of spatial variability of soil P within a field revealed that, VRP should be implemented over small areas (perhaps down to few square meters). This can be achieved with an on-the-go P fertilisation system. For on-the-go P fertilisation, three important elements are required: a soil sensor to gather soil information while travelling across a field, a model to derive the actual soil P and a recommendation model to predict corresponding application rate. A VIS-NIR soil sensor for on-the-go gathering of soil spectra (Mouazen et al, 2005) was installed at the front of a planter-applicator for on-the-go gathering of soil spectra. The optical sensor unit to measure soil reflectance spectra was attached to a mobile, fibre-type, VIS-NIR spectrophotometer (CORONA FIBRE VISNIR 1.7, Carl Zeiss) with a measurement range of 305 - 1711 nm. Two visible (VIS) and near infrared (NIR) models were developed for prediction of soil P using wet soils based on soil samples that were collected from Belgium and Northern France. These models were calibrated based on two different measures of soil available P: Olsen P and ammonium lactate extractable P (Pal). The performance of the models was validated by different methodologies. The accuracy of the model was verified using several independent field samples. Since in Belgium the phosphate is recommended based on Pal test, the model developed (Mouazen et al, 2007) based on the Pal was used as basis for VR fertilisation system developed for this dissertation. Another challenge to on-the-go P application is fertiliser recommendation. Generally, fertiliser recommendation is based on many factors rather than just measured soil P. These include soil type, pH, other soil nutrients and soil moisture content. These soil parameters must be simultaneously measured in addition to soil P for actual fertiliser recommendation. Since, for the present research these factors were not available in on-the-go measurements, a recommendation model was established solely based on soil Pal measure. This simplified model was derived from the information provided by Soil Service of Belgium rather than based on the Expert System (Vandendriessche et al., 1993 & 1995) for precise fertiliser recommendation that uses many of the above mentioned parameters. The fertiliser recommendation based on this simplified model could sometimes fail due to ignoring other factors. But this could be remedied in future. A programme (LabVIEW, National Instrument) was developed to record soil spectra, predict soil Pal during on-the-go measurement and calculate required P fertiliser amount. This algorithm provided the signal to the VR fertiliser applicator (ED302, AMAZONE) to change the application rate accordingly. In addition, the coordinates of the sampling points were recorded using a DGPS (Trimble® AgDGPS 132) that was included to document the soil and fertiliser maps, although, for a sensor-based VR application system this positioning system is not required. In an experimental field alternative plots were selected so that every other plot was used for VR application as for uniform rate (UR) treatment. Maize was planted over the entire field. During the planting, only the plots assigned for VR treatment received the rates recommended by on-the-go measurement and the rest received a uniform dose of 30 kg P2O5/ha based on the standard Pal measurement and recommendation by the Soil Service of Belgium. The location of the collected soil P spectra and the corresponding phosphate application rate was recorded using a DGPS. The number of plant leaves and grain yield were considered as growth indices, which might be influenced by P deficiency. The maize was harvested using a combine harvester (CR 960 New Holland) instrumented with a continuous grain flow sensor (New Holland) and a DGPS, which allowed positioning the grain yield while travelling across the field. The plant leaves and crop yield showed lower variation in plots assigned for VR treatment compared to those of UR treatment (coefficient of variation of 25 and 31%, respectively). This may be due to the proper distribution of the fertiliser according to soil requirement. No significant difference was found for the number of plant leaves between VR and UR plots. However, the yield of the plots with the VR treatment was significantly higher than for the UR treatment. Yield increase of 336 kg grain/ha was observed in the VR plots. However, the experimental layout of the plots could not be properly designed using a statistical randomisation method because of the limitation imposed by the machine operation. Therefore, this result needs further investigation to be confirmed. Average application rate of phosphate on VR plots was 28.75 kg P2O5/ha which is 1.25 kg P2O5/ha lower than the recommended UR (30 kg P2O5/ha). This implies that the benefit of using VR strategy is minor in agricultural fields with a high level of soil Pal which is the case for most of the Belgian fields.

Listing 1 - 1 of 1
Sort by