Factorial Designs

This chapter moves on from simple ‘one-way’ designs to more complex factorial designs. It extends the simple linear model to include interactions as well as average main effects. Interactions are assessed relative to a null additive expectation where the treatments have no effect on each other. Interactions can be positive, when effects are more than additive, or negative, when they are less than expected. The chapter considers in detail the analysis of an example data set concerning the mechanisms of loss of plant diversity following fertilizer treatment.

Crown projection area of Licania tomentosa (Benth.) Fritsch (Chrysobalanaceae), estimated by linear regression / Área de projeção da copa de Licania tomentosa (Benth.) Fritsch (Chrysobalanaceae), estimada por regressão linear

Knowledge of the Crown Projection Area (CPA) allows to make inferences about the shading and to know space occupied by a tree. However, crown measurements are more time-consuming and laborious when compared to those of Circumference Breast Height (CBH). Thus, this work aimed to evaluate regression models and present the most suitable to CPA estimate of Licania tomentosa, in an urban area of São João Evangelista municipality, Brazil. Fifty trees distributed over 7 public roads were sampled. CBH and Crown Diameter (CD, m) were measured for later calculation of its projection area (CPA, m2). Four regression models were tested in order to estimate CPA as a function of CBH alone. The equation derived from of the model “” showed a homoscedastic distribution of the percentage residues, with closer deviations around the abscissa axis. It is concluded that the equation obtained with the adjustment of the simple linear model was the most efficient to estimate of the crown projection area of L. tomentosa. This projection area increased as the stem of the trees thickened.

Honey bees communicate distance via non-linear waggle duration functions

Honey bees (genus Apis) can communicate the approximate location of a resource to their nestmates via the waggle dance. The distance to a goal is encoded by the duration of the waggle phase of the dance, but the precise shape of this distance-duration relationship is ambiguous: earlier studies (before the 1990s) proposed that it is non-linear, with the increase in waggle duration flattening with distance, while more recent studies suggested that it follows a simple linear function (i.e. a straight line). Strikingly, authors of earlier studies trained bees to much longer distances than authors of more recent studies, but unfortunately they usually measured the duration of dance circuits (waggle phase plus return phase of the dance), which is only a correlate of the bees’ distance signal. We trained honey bees (A. mellifera carnica) to visit sugar feeders over a relatively long array of distances between 0.1 and 1.7 km from the hive and measured the duration of both the waggle phase and the return phase of their dances from video recordings. The distance-related increase in waggle duration was better described by a non-linear model with a decreasing slope than by a simple linear model. The relationship was equally well captured by a model with two linear segments separated at a “break-point” at 1 km distance. In turn, the relationship between return phase duration and distance was sufficiently well described by a simple linear model. The data suggest that honey bees process flight distance differently before and beyond a certain threshold distance. While the physiological and evolutionary causes of this behavior remain to be explored, our results can be applied to improve the estimation of honey bee foraging distances based on the decoding of waggle dances.

Mapping the suitability of groundwater-dependent vegetation in a semi-arid Mediterranean area

Mapping the suitability of groundwater-dependent vegetation in semi-arid Mediterranean areas is fundamental for the sustainable management of groundwater resources and groundwater-dependent ecosystems (GDEs) under the risks of climate change scenarios. For the present study the distribution of deep-rooted woody species in southern Portugal was modeled using climatic, hydrological and topographic environmental variables. To do so, Quercus suber, Quercus ilex and Pinus pinea were used as proxy species to represent the groundwater-dependent vegetation (GDV). Model fitting was performed between the proxy species Kernel density and the selected environmental predictors using (1) a simple linear model and (2) a geographically weighted regression (GWR) to account for autocorrelation of the spatial data and residuals. When comparing the results of both models, the GWR modeling results showed improved goodness of fit as opposed to the simple linear model. Climatic indices were the main drivers of GDV density, followed by a much lower influence by groundwater depth, drainage density and slope. Groundwater depth did not appear to be as pertinent in the model as initially expected, accounting only for about 7 % of the total variation compared to 88 % for climate drivers. The relative proportion of model predictor coefficients was used as weighting factors for multicriteria analysis to create a suitability map for the GDV in southern Portugal showing where the vegetation most likely relies on groundwater to cope with aridity. A validation of the resulting map was performed using independent data of the normalized difference water index (NDWI), a satellite-derived vegetation index. June, July and August of 2005 NDWI anomalies, for the years 1999–2009, were calculated to assess the response of active woody species in the region after an extreme drought. The results from the NDWI anomalies provided an overall good agreement with the suitability to host GDV. The model was considered to be reliable for predicting the distribution of the studied vegetation. The methodology developed to map GDVs will allow for the prediction of the evolution of the distribution of GDV according to climate change and aid stakeholder decision-making concerning priority areas of water resource management.

Sentiment Analysis - Interview Transcripts

One-on-one interviews and the analysis of their transcripts is the most common way researchers get into depth to obtain detailed insights. These insights are highly subjective and lack objectivity. We demonstrate in this paper a method and a use case to bring objectivity to this analysis. We present the use of NLP to generate sentiment analysis and perform various quantitative techniques. This analysis is useful in deriving insights by finding patterns and building a simple linear model to explain the variation in sentiment pattern. We also present a view about the usage of this technique for the effective and optimal time usage by researchers to learn maximum from outlier interviews.

Mapping the suitability of groundwater dependent vegetation in a semi-arid Mediterranean area

The forecasted groundwater resource depletion under future climatic conditions will greatly influence subsurface groundwater dependent ecosystems and their associated vegetation. In the Mediterranean region this will create harsh conditions for the maintenance of agroforestry systems dependent on groundwater, such as cork oak woodlands. The threat of increasing aridity conditions will affect their productivity and eventually induce a shift in their geographical distribution. Thus, characterizing and modelling the relationship between environmental conditions and subsurface groundwater dependent vegetation (subsurface GDV) will allow to identify the main drivers controlling its distribution and predict future impacts of climate change. In this study, we built a model that explains subsurface GDV distribution in southern Portugal from climatic, hydrological and topographic environmental variables. To achieve this, we relied on the density of Quercus suber, Quercus ilex and Pinus pinea as proxy species of subsurface GDV. Model fitting was performed between the proxy species Kernel density and the selected environmental predictors using (1) a simple linear model and (2) a Geographically Weighted Regression (GWR), to account for auto-correlation of the spatial data and residuals. When comparing the results of both models, the GWR modelling results showed improved goodness of fitting, as opposed to the simple linear model. Soil type was the main driver of subsurface GDV density closely followed by the aridity index. Groundwater depth did not appear to be as pertinent in the model as initially expected. Model predictor coefficients were used as weighting factors for multicriteria analysis, to create a suitability map to the subsurface GDV in southern Portugal. A validation of the resulting map was performed using independent data of integrated potential distribution of each proxy tree species in the region and overall, there was an accordance between areas of good suitability to subsurface GDV. The model was considered reliable to predict the distribution of the studied vegetation, however, lack of data quality and information was shown to be the main cause for suitability discrepancies between maps. Our new methodology on mapping of subsurface GDV's will allow to predict the evolution of the distribution of subsurface GDV according to climate change scenarios and aid stakeholder decision-making concerning priority areas of water resources management.