Using remote sensing in support of environmental management: A framework for selecting products, algorithms and methods.

Traditionally, to map environmental features using remote sensing, practitioners will use training data to develop models on various satellite data sets using a number of classification approaches and use test data to select a single 'best performer' from which the final map is made. We use a combination of an omission/commission plot to evaluate various results and compile a probability map based on consistently strong performing models across a range of standard accuracy measures. We suggest that this easy-to-use approach can be applied in any study using remote sensing to map natural features for management action. We demonstrate this approach using optical remote sensing products of different spatial and spectral resolution to map the endemic and threatened flora of quartz patches in the Knersvlakte, South Africa. Quartz patches can be mapped using either SPOT 5 (used due to its relatively fine spatial resolution) or Landsat8 imagery (used because it is freely accessible and has higher spectral resolution). Of the variety of classification algorithms available, we tested maximum likelihood and support vector machine, and applied these to raw spectral data, the first three PCA summaries of the data, and the standard normalised difference vegetation index. We found that there is no 'one size fits all' solution to the choice of a 'best fit' model (i.e. combination of classification algorithm or data sets), which is in agreement with the literature that classifier performance will vary with data properties. We feel this lends support to our suggestion that rather than the identification of a 'single best' model and a map based on this result alone, a probability map based on the range of consistently top performing models provides a rigorous solution to environmental mapping.

Spatiotemporal analysis of encroachment on wetlands: a case of Nakivubo wetland in Kampala, Uganda.

Wetlands provide vital ecosystem services such as water purification, flood control, and climate moderation among others, which enhance environmental quality, promote public health, and contribute to risk reduction. The biggest threat to wetlands is posed by human activities which transform wetlands, often for short-term consumptive benefits. This paper aimed to classify and map recent land cover and provide a multi-temporal analysis of changes from 2002 to 2014 in the Nakivubo wetland through which wastewater from Kampala city drains to Lake Victoria in Uganda. The paper contributes through spatially congruent change maps showing site-specific land cover conversions. In addition, it gives insight into what happened to the wetlands, why it happened, how the changes in the wetlands affect the communities living in them, and how the situation could be better managed or regulated in future. The analysis is based on very high resolution (50-62 cm) aerial photos and satellite imagery, focus group discussions, and key informant interviews. Overall, the analysis of losses and gains showed a 62 % loss of wetland vegetation between 2002 and 2014, mostly attributable to crop cultivation. Cultivation in the wetland buffering the lake shore makes it unstable to anchor. The 2014 data shows large portions of the wetland calved away by receding lake waves. With barely no wetland vegetation buffer around the lake, the heavily polluted wastewater streams will lower the quality of lake water. Furthermore, with increased human activities in the wetland, exposure to flooding and pollution will be likely to have a greater impact on the health and livelihoods of vulnerable communities. This calls for a multi-faceted approach, coordination of the various stakeholders and engagement of wetland-dependent communities as part of the solution, and might require zoning out the wetland and restricting certain activities to specific zones.

Is acarbose equivalent to tolbutamide as first treatment for newly diagnosed type 2 diabetes in general practice? A randomised controlled trial.

We performed a double blind randomised controlled trial in general practice to assess equivalence between tolbutamide and acarbose with respect to the effect on mean HbA(1c) in newly diagnosed patients with type 2 diabetes. Secondary objectives were to compare the effects of both treatments on fasting and post-load blood glucose and insulin levels, lipids, and adverse events. Patients were randomised to receive acarbose, titrated step-wise to a maximum of 100mg three times daily (n=48) or tolbutamide, similarly titrated to a maximum of 2000 mg in three doses (n=48). The two treatments were considered equivalent if the two-sided 90% confidence interval (CI) for the difference in mean HbA(1c) levels was within the range -0.4 to 0.4%. Results were analysed on an intention-to-treat, per-protocol and on worst-case basis. Both agents reduced the HbA(1c) percentage and fasting blood glucose levels. The difference in mean decrease of HbA(1c) was 0.6% in favour of tolbutamide (90% CI 0.3, 0.9; 95% CI 0.2, 1.0). A worst-case analysis, assuming no change in HbA(1c) for dropouts, yielded a difference in mean decrease of 0.9% (90% CI 0.6, 1.2) in favour of tolbutamide. The difference in mean decrease of fasting blood glucose was 1.0 mmol/l in favour of tolbutamide (95% CI 0.3, 1.7). There were no significant differences in post-load blood glucose, fasting and post-load insulin levels, or lipids. In the acarbose group significantly more patients (15 versus 3) discontinued therapy because of adverse effects, mostly of gastrointestinal origin. We conclude that the results of this study favour tolbutamide over acarbose as first treatment for patients with newly diagnosed type 2 diabetes.