Possibilities and limitations of artificial neural networks for subpixel mapping of land cover

Although developments in remote sensing have greatly improved land cover mapping, the mixed pixel problem has not yet been fully addressed. Soft classification techniques have been introduced to address the problem, but they do not show the spatial location of the class proportions in a pixel. Subpixel mapping has been introduced to address the drawbacks of soft classifications. In this work, the feedforward backpropagating neural network (FFBPNN) was used for subpixel mapping. A set of class proportion images, which are to be treated as soft classification results, were created from a high spatial resolution (25 m) land cover dataset. For this purpose, the land cover dataset was aggregated both thematically (into two, four or eight land cover classes) and spatially (into proportion images with pixel sizes of 75, 150 and 300 m). This resulted in nine different combinations that were considered here as study cases. Several FFBPNNs were trained using these proportion images and the original land cover dataset (which was used as a target). Subsequently, the best networks were used to reconstruct high spatial resolution land cover maps of two heterogeneous areas in the south of The Netherlands. The overall accuracies obtained revealed that the networks were influenced by the spatial frequency, shape and size of the different land cover types. Moreover, it was revealed that most of the errors were on the class boundaries where highly mixed pixels are to be expected. The accuracies spanned a wide range of values depending on the complexity of the cases. Although it was not possible to exhaustively explore all network architectures, the results demonstrate the potential of the FFBPNN for subpixel mapping.     

Model extraction of micromachined planar inductors implemented on alumina substrate

In this work, the high resistivity material property of Alumina substrate is made into use in order to minimize the substrate loss and thereby rendering of relatively high quality MEMS suspended inductors. The modified \(\pi \)-model is used in order to account the skin effect as well as exploit the frequency-dependent losses in the devices. On-wafer-measurements were made prior to characterization and model extraction. The maximum quality factor achieved is about 37.2 with inductor value 3.98 nH at 1.28 GHz. This result is quite promising and could be further enhanced by increasing the metal thickness.     

Frequent Absence of GBSS 1 B Isoprotein in Endosperm Starch of Canadian Wheat Cultivars

Granule bound starch synthase 1 (GBSS 1) is an enzyme involved in amylose synthesis, and it is encoded at the waxy loci in wheat. Mutations have been reported for three GBSS1 loci in wheat. Mutations at wheat GBSS 1 loci result in the absence of GBSS 1 isoproteins with concomitant reduction in the amylose concentrations of endosperm starch. A recent survey of Canadian wheat cultivars detected the frequent absence of GBSS1 B isoprotein in endosperm starch granules. This was likely a result of many Canadian spring wheat cultivars having common parents, which have a mutation at the GBSS1 B locus.     

Comparative study of stationary phases impregnated with acidic organophosphorus extractants for HPLC separation of rare earth elements

Chromatographic separation behavior of rare earth elements (REEs) on reversed-phase columns impregnated with di-(2-ethylhexyl) phosphoric acid (H[DEHP]), 2-ethylhexyl, 2-ethylhexyl phosphonic acid (H[(EH)EHP]) and di-(2,4,4-trimethylpentyl) phosphinic acid (H[TMPeP]) were compared utilizing nitric acid as eluent and ICP-MS for detection. Retention factors were determined and compared under the same eluting conditions using columns loaded with approximately 0.4 mmole of each extractant. The separation efficiency was evaluated from resolution data of the REEs acquired under same and optimized elution conditions for each column. The best separation for the entire series of REEs was obtained with H[(EH)EHP], H[DEHP] was found more appropriate for lighter REEs and H[TMPeP] was the best choice for separation of heavy REEs.     

Species-specific PCR-based assays for the detection of Fusarium species and a comparison with the whole seed agar plate method and trichothecene analysis

Species-specific PCR was used for the identification of nine Fusarium species in pure mycelial culture. A PCR-based method was compared with the whole seed agar plate method and trichothecene analysis for three toxin-producing Fusarium species using 85 grain samples of wheat, barley, oat, corn and rye. A simple SDS-based DNA extraction system followed by potassium acetate precipitation resulted in consistent PCR amplification of DNA fragments from cultures and grain samples. The species-specific PCR assays correctly identified pure cultures of Fusarium avenaceum ssp. avenaceum (9 isolates), Fusarium acuminatum ssp. acuminatum (12 isolates), Fusarium crookwellense (7 isolates), Fusarium culmorum (12 isolates), Fusarium equiseti (11 isolates), Fusarium graminearum (77 isolates), Fusarium poae (10 isolates), Fusarium pseudograminearum (23 isolates), and Fusarium sporotrichioides (10 isolates). Multiplex PCR was developed for the simultaneous detection of F. culmorum, F. graminearum and F. sporotrichioides, the three most important trichothecene producing species in Canada. In grain samples, results of PCR assays for these same three species related well with whole seed agar plate method results and determination of Fusarium trichothecenes. The PCR assay described in this study can be used for routine detection and identification of Fusarium spp. in Canada.     

Phytoplankton composition and biomass in tropical soda Lake Shala: seasonal changes in response to environmental drivers

Although soda lakes are valuable, sensitive aquatic resources where phytoplankton play a decisive role for the entire ecological functions, they are among the least‐studied ecosystems. Seasonal variations in phytoplankton composition, abundance and biomass in relation to some environmental parameters of the little known, deep, large, volcanic and saline–alkaline Lake Shala were investigated over an annual cycle. The lake phytoplankton community consisted of relatively diverse taxa (23) belonging to Bacillariophyceae, Cryptophyta, Cyanoprokaryota and Dinophyta. Bacillariophyceae and Cryptophyta were the dominant groups throughout the annual cycle, accounting for about 57% and 22% of the total number of species, and 28% and 69% of the total abundance of the phytoplankton community, respectively. Cryptomonas spp. were most abundant throughout nearly all months, contributing about 59%–95% of total phytoplankton abundance, followed by Thalassiosira sp. (1%–35%). The chlorophyll‐a concentration, as a proxy for algal biomass, was generally low (mean 17 μg L^?1), exhibiting only small seasonal variation. The strong, inverse relation of chlorophyll‐a with water transparency (r = ?0.69; n = 11) and the persistent dominance of species adapted to low‐light conditions and mixing suggest the overriding importance of these factors in controlling the lake's phytoplankton. The results of the present study generally suggest the phytoplankton composition and biomass in Lake Shala exhibited muted seasonal changes, despite the environmental perturbations, probably because of the lake's high buffering capacity against allochthonous impacts because of its voluminous nature.     

Estimating sediment yield risk of reservoirs in northern Ethiopia using expert knowledge and semi‐quantitative approaches

Reservoir siltation is a serious problem that threatens the productivity and sustainability of water‐harvesting schemes. Quantification of sediment deposition in reservoirs and understanding of its major drivers are needed to apply targeted management interventions. Most of the techniques used to estimate sediment deposition in reservoirs require extensive measurements on a frequent time basis, as well as being costly and time‐consuming. Thus, a rapid and relatively economical means of assessing the erosion susceptibility of catchments and predicting their sediment yield potential is necessary. In this study, expert‐based rankings and semi‐quantitative factorial scoring approaches were applied to assess the siltation severity of 25 reservoirs in the Tigray region of northern Ethiopia. The results were then correlated with quantitative sediment yield estimates acquired for representative sites, and a sediment yield predication model was developed for the region. The calibrated model has an efficiency and relative root mean square error (RRMSE) of 79 and 36%, respectively, which is considered adequate to assess erosion susceptibility and siltation risk of reservoirs in similar environments. The study demonstrates that expert knowledge and rapid characterization of catchments, in terms of susceptibility to erosion, are viable options for assessing siltation risks and for analysing controlling factors at a larger number of sites, with minimum costs and acceptable accuracy.