The Effect of Changes to the Method of Estimating the Pollen Count from Aerobiological Samples

Pollen data have been recorded at Novi Sad in Serbia since 2000. The adopted method of producing pollen counts has been the use of five longitudinal transects that examine 19.64% of total sample surface. However, counting five transects is time consuming and so the main objective of this study is to investigate whether reducing the number to three or even two transects would have a significant effect on daily average and bi-hourly pollen concentrations, as well as the main characteristics of the pollen season and long-term trends. This study has shown that there is a loss of accuracy in daily average and bi-hourly pollen concentrations (an increase in % ERROR) as the sub-sampling area is reduced from five to three or two longitudinal transects. However, this loss of accuracy does not impact on the main characteristics of the season or long-term trends. As a result, this study can be used to justify changing the sub-sampling method used at Novi Sad from five to three longitudinal transects. The use of two longitudinal transects has been ruled out because, although quicker, the counts produced: (a) had the greatest amount of % ERROR, (b) altered the amount of influence of the independent variable on the dependent variable (the slope in regression analysis) and (c) the total sampled surface (7.86%) was less than the minimum requirement recommended by the European Aerobiology Society working group on Quality Control (at least 10% of total slide area).

Defining Spatial and Temporal Hydromorphological Sampling Strategies for the Leigh Brook River Site

Detailed surveys of depth and velocity are undertaken to describe hydro-ecological status of rivers. Fieldwork for these surveys is time consuming and expensive. This paper aims to describe the methodology applied in order to determine the most suitable depth sampling strategy for effective field data collection and river representation in time and space at the Leigh Brook river site, Worcester, UK. The accuracy of three different sampling strategies for predicting depth at non-measured points has been compared and the mesohabitats that better characterise depth changes due to variations in discharge have been identified. The results show that depth changes due to discharge change are mainly located at shallow and deep glide mesohabitat types. The analysis for the comparison of sampling strategies indicates that grid sampling strategies give better results than regular transects. Since the results also show that higher errors in predictions are obtained in the deepest areas, higher sampling densities should be applied in these locations.