Thesis
African eggplant (AEP) (Solanum aethiopicum L) is one of the most consumed
vegetables in Tanzania. The crop is important for nutritional and food security purposes
as well as income generation for smallholder farmers. The crop is also rich in potassium,
magnesium, calcium and iron (Hossein et al., 2010). It is also a potential genetic resource
for breeding of eggplant and grafting rootstock of hybrid eggplant that improve its
productivity. This is due to its tolerance to diseases as well as biotic and abiotic stresses
(Sabatino et al. 2018; 2019).
Despite its benefits, AEP has been neglected in terms of climate adaptation and
sustainable agricultural intensification in East Africa. This has led to reduced levels of
yields from
42 – 57 t/ha to 15 t/ha (Fondio et al., 2016; Msogoya et al., 2014;
Shackleton et al., 2009). The yield gap is due to inadequate water and nitrogen (N)
management practices during the growing season among other factors (Chaves et al.,
2010). It thus requires optimization of water and nitrogen management using most
efficient and effective methodologies which will ensure water and N use efficiencies to
boost yields as well as protect the environment.
The study was therefore conducted to examine the applicability of remote sensing
technology for optimization of water and nitrogen management in African eggplant
(Solanum aethiopicum L) production under irrigation in tropical sub-humid areas. The
main objective of the study was to assess the unmanned aerial vehicles (UAVs)
multispectral imaging remote sensing technology for optimization of water and nitrogen
management in irrigated African eggplant production under tropical sub-humid
conditions. The specific objectives of the study were (1) to evaluate the optimum
interaction of water and nitrogen application for African eggplant, (2) to investigate the
feasibility of hand-held thermal and multispectral imaging in managing of water use for
African eggplant, (3) to examine the efficacy of un-manned aerial vehicles vegetation
indices to depict nitrogen in irrigated African eggplant, (4) to investigate the applicability
of UAV-based multispectral vegetation indices for managing the interaction between
water and nitrogen in irrigated African eggplant. The experimental study composed two types of layout based on the specific objectives.
To study the effects of water and nitrogen independently, the study was conducted in a
randomized block design (RBD). To assess the interactive effects of water and nitrogen
application the experiment was laid-out in a randomized split-plot design with irrigation
being the main and nitrogen (N) treatments as a sub-factor. The irrigation regimes were
100% (I100), 80% (I80) and 60% (I60) of crop water requirements whilst nitrogen levels
were 250 kg N/ha (F100), 187 kg N/ha (F75), 125 kg N/ha (F50) and 0 kg N/ha (F0).
Mobile phone-based thermal images were used to collect data for the crop water stress
index (CWSI). The orthomosaic images were acquired using a multispec4c sensor
attached in the UAV. UAV images rectification and mosaicking was conducted through
PIX4D mapper software (Pix4D SA, Lausanne, Switzerland). The image mosaics were
stored in separate bands (green, red, red edge and near-infrared). The stored orthomosaic
maps were trimmed to the field level using the extraction clipper application in the QGIS
2.18 software.
A geometric correction process (image to image
registration) was conducted to ensure that images for each of the treatments corresponded
to the same location throughout the irrigation cycle (Dave et al., 2015). Images in
different spectral bands were extracted using a python script and used to compute
different vegetation indices through raster and rgdal packages in R software (version
3.6.1).
The results have shown that, water and N have an influence on crop growth variables
(plant height and LAI) consequently yield increase. Both plant height and leaf area index
(LAI) had a good correlation with fruit yield (R 2 = 0.6 and 0.8). The combination of
irrigation regime of 100% and 75% N yielded the best performance in fruit quality. The
best water use efficiency (WUE) and agronomic nitrogen use efficiency (NUE) was
attained at 80% and 100% levels of water in combination with 75% N. However, taking
into consideration different factors the optimum water and nitrogen application was 80%
and 75% of the total irrigation and N requirements.
The handheld-thermal imaging and unmanned aerial vehicles (UAVs) in water
management have shown different capabilities. The crop water stress index (CWSI)
derived from the mobile phone-based thermal images was sensitive to leaf moisture
content (LMC) in I80 and I60 at all vegetative stages. The UAV-derived Normalized
Difference Vegetation Index (NDVI) and Optimized Soil Adjusted Vegetation Index (OSAVI) correlated with LMC at the vegetative and full vegetative stages for all three
irrigation treatments.
Therefore, where eggplant is grown under deficit irrigation, CWSI can be used together
with to NDVI or OSAVI at vegetative or full vegetative stages depending on available
resources. Similarly, for nitrogen (N) management under tropical sub-humid conditions,
the study recommended the application of green normalized difference vegetation index
(GNDVI), NDVI and OSAVI as they best determined leaf N concentration (p<0.01).
Furthermore, none of the vegetation index was capable to clearly distinguish the
interactive effects of water and N. The vegetation indices were more sensitive to
distinguish water and N separately. For instance, the transformed difference vegetation
index (TDVI) was more sensitive to water variation during the vegetative stage of crop
development.
The study therefore concluded that, the UAV based multispec4c sensor, has capability to
be used in optimizing water and nitrogen in drip system irrigated African eggplant, grown
under tropical sub-humid conditions areas.
Feed the Future Innovation Lab for Small Scale Irrigation (ILSSI) through the U.S. Agency for International Development and CGIAR Program on Water, Land, and Ecosystems (WLE)