Predicting tree species presence and basal area in Utah: A comparison of stochastic gradient boosting,generalized additive models,and tree-based methods

Many efforts are underway to produce broad-scale forest attribute maps by modelling forest class and structure variables collected in forest inventories as functions of satellite-based and biophysical information. Typically, variants of classification and regression trees implemented in Rulequest's^© See5 and Cubist (for binary and continuous responses, respectively) are the tools of choice in many of these applications. These tools are widely used in large remote sensing applications, but are not easily interpretable, do not have ties with survey estimation methods, and use proprietary unpublished algorithms. Consequently, three alternative modelling techniques were compared for mapping presence and basal area of 13 species located in the mountain ranges of Utah, USA. The modelling techniques compared included the widely used See5/Cubist, generalized additive models (GAMs), and stochastic gradient boosting (SGB). Model performance was evaluated using independent test data sets. Evaluation criteria for mapping species presence included specificity, sensitivity, Kappa, and area under the curve (AUC). Evaluation criteria for the continuous basal area variables included correlation and relative mean squared error. For predicting species presence (setting thresholds to maximize Kappa), SGB had higher values for the majority of the species for specificity and Kappa, while GAMs had higher values for the majority of the species for sensitivity. In evaluating resultant AUC values, GAM and/or SGB models had significantly better results than the See5 models where significant differences could be detected between models. For nine out of 13 species, basal area prediction results for all modelling techniques were poor (correlations less than 0.5 and relative mean squared errors greater than 0.8), but SGB provided the most stable predictions in these instances. SGB and Cubist performed equally well for modelling basal area for three species with moderate prediction success, while all three modelling tools produced comparably good predictions (correlation of 0.68 and relative mean squared error of 0.56) for one species.     

Organic carbon stratification and size distribution of three typical paddy soils from Taihu Lake region, China

Developing realistic soil carbon (C) sequestration strategies for China’s sustainable agriculture relies on accurate estimates of the amount, retention and turnover rates of C stored in paddy soils. Available C estimates to date are predominantly for the tilled and flood-irrigated surface topsoil (ca. 30 cm). Such estimates cannot be used to extrapolate to soil depths of 100 cm since soil organic carbon (SOC) generally shows a sharp decrease with depth. In this research, composite soil samples were collected at several depths to 100 cm from three representative paddy soils in the Taihu Lake region, China. Soil organic carbon distribution in the profiles and in aggregate-size fractions was determined. Results showed that while SOC decreased exponentially with depth to 100 cm, a substantial proportion of the total SOC (30%–40%) is stored below the 30 cm depth. In the carbon-enriched paddy topsoils, SOC was found to accumulate preferentially in the 2–0.25 and 0.25–0.02 mm aggregate size fractions. σ13C analysis of the coarse micro-aggregate fraction showed that the high degree of C stratification in the paddy topsoil was in agreement with the occurrence of lighter @1313C in the upper 30 cm depth. These results suggest that SOC stratification within profiles varies with di erent pedogenetical types of paddy soils with regards to clay and iron oxyhydrates distributions. Sand-sized fractions of aggregates in paddy soil systems may play a very important role in carbon sequestration and turnover, dissimilar to other studied agricultural systems.     

Regionale Klimatrends mit Hilfe der phänologischen Uhr, dargestellt am Beispiel Rheinland-Pfalz

Aims and Scope

Over recent decades, the climate has changed. Up to now this statement has been made predominantly for the global and/or continental area. What, however, is the situation regarding the climatic trend at the regional scale? Is it already possible to show effects of a climate change in small-scale landscape structures? These questions were addressed looking at the example of the situation in the federal state of Rhineland-Palatinate (Germany).

Methods

Phenological data of plants of the last 50 years from a total of 386 stations of the German Weather Service (DWD) were evaluated. The timing of the different plant phases, beginning of the foliation and plant senescence in autumn, flowering and the first ripening of the fruits had to be registered for different agricultural crops, wild plants, ornamental woods as well as fruit trees. The analysed data material was used to develop the threefold phenological clock (DPU).

Results

Summarising the findings of the DPU for the 15 evaluated nature districts of Rhineland-Palatinate according to individual seasons, which were defined in generic terms as spring, summer, autumn and winter, a tendency of a prematuring of the spring, a prolongation of the autumn as well as a shortening of the winter is evident. From a regional point of view, on the other hand, in individual nature districts such as the ‘Westerwald’ and the ‘Bergisch-Sauerländische Gebirge’, winter seasons are getting longer. Thus the effects of a climate change can differ considerably from region to region.

Outlook

The regionally quite different changes of the climatic situation in Rhineland-Palatinate will not remain without an effect on the ecosystems of this federal state. The currently existing biocoenosis will therefore change. The consequences for forestry, agriculture and nature conservation are of special interest in this context.     

Brood defence in the great tit (Parus major): the influence of life-history and habitat

Summary We examined the extent to which parental investment, as measured by brood defence, is determined by life-historical selection in a shortlived bird, the great tit (Parus major). Pairs tending first (n=20) and second (n=21) broods in the same Scots pine woods in 1983 were used to test predictions of a cost/benefit model of brood defence based on the species' average demography in coniferous forest. Furthermore, the differences in demography between pine and deciduous forest permitted us to test whether habitat-specific life-history would affect the seasonal pattern of defence. In the model, benefit was defined as the brood's potential contribution to a parent's fitness, and the cost as the potential loss if the defender dies in the act of defence. Univariate and multivariate procedures were applied to six measures of defence response to a live owl (Glaucidium perlatum) plus a taped mobbing chorus. Results proved three of the model's predictions to be false. In coniferous forest, neither the overall strength nor the individual variance of defence behaviour differed among first and second broods, nor was there any consistent difference in defence strength between pairs living in coniferous (n=54) or deciduous (n=84) forest as revealed by comparisons within each brood. These failures could be reconciled with the model by assuming that selection in the past had acted via the average demography of both types of habitat. The model received direct support from defence strength increasing with age of young and, more forcefully, becoming more influenced by brood size in second broods, regardless of habitat.A difference in the strength of defence by the male and female suggests two more functions of behaviour: In first broods, the male risks more than the female as measured by five of six variables. This suggests that defence is facultatively linked to the need for territorial protection from predators all the year round. In the female's presence, the male, taking an additional risk, approaches the owl to half the distance of that of a single, though paired, male, suggesting an additional, social role of defence behaviour.Taken together, anti-predator defence in the great tit serves to protect the brood, the home range and, in the male, the female mate. The magnitude of the benefits envisaged varies among the sexes.     

The effect of 4-Nonylphenol on the pigmentation of<Emphasis Type="Italic">Ocimum basilicum</Emphasis> (Basil)

BACKGROUND, AIM AND SCOPE: Tests during the last few years have confirmed that 4-nonylphenol (4-NP) can have oestrogen-like effects (xeno-hormone) on animal organisms. The objective was to firstly evaluate the ecotoxicological effects of 4-NP on plants such as hydrocultures. To clarify how this substance interferes with the photosynthetic system of plants, various tests were carried out using the basil plant (Ocimum basilicum). METHODS: The effect of the pollutant 4-NP on the pigment content in the leaves of the basil plant was analysed with the use of High-Performance-Liquid-Chromatography (HPLC). RESULTS AND DISCUSSION: A general assessment of the HPLC data revealed that plants that came in contact with the 4-nonylphenol showed a change in pigmentation. More chlorophyll a and b was produced, although at the same time a higher production of degradation products and by-products of the chlorophylls was observed. These occurrences can therefore be seen as an impairment of the photosynthetic process. The contaminated plants produced less xanthophylls than the non-contaminated ones, though these differences were statistically not significant. CONCLUSION AND PERSPECTIVE: The variations on the pigment content in the leaves of the basil plant can be interpreted as a consequence of the 4-NP application. It was, however, not investigated whether the plants absorbed the pollutants directly. The effect could have been caused by adsorption of the oily substance to the roots, and this could have led to a hindrance of the uptake of nutrients and possibly water. In order to clarify this further, biochemical experiments are being conducted.     

In vivo laser-induced fluorescence detection of pyrene in nematodes and determination of pyrene binding constants for humic substances by fluorescence quenching and bioconcentration experiments

The bioconcentration of pyrene by bacterivorous thread worms (nematodes) of the species Caenorhabditis elegans was studied with laser-induced fluorescence (LIF) spectroscopy, fluorescence imaging and a radiotracer method. The vibronic band intensities of the LIF spectra indicated that the microenvironment of pyrene in the nematodes was similar to a low-polarity solvent, and thus provided direct evidence that pyrene was accumulated in lipid-rich areas inside the nematodes. The concentration of pyrene in the nematodes was estimated from the monomer/excimer fluorescence intensity ratio. Results from this method were in fair agreement with results using 14C labeled pyrene for measuring pyrene bioconcentration. Preliminary results indicated that LIF measurements of pyrene may be possible even in single nematodes. Fluorescence microscopic observations revealed that pyrene was not adsorbed on the outside of the organisms, but was strongly concentrated in restricted areas inside the worms. In the second part of the study, the effects of six different humic substances (HS) on the bioconcentration of pyrene were investigated and sorption coefficients (KDOC) calculated from reductions in bioconcentration (KDOC(biol)) were compared with sorption coefficients measured with a fluorescence quenching technique (KDOC(flu)). The results of these two different experimental methods agreed well (with KDOC(biol) being slightly lower than KDOC(flu), indicating that the fraction of pyrene that was determined as freely dissolved by the fluorescence quenching method was comparable to the bioavailable fraction.     

Besprechungen

The Science of Nature -     

Occupational exposure in the fluorescent lamp recycling sector in France

The fluorescent lamp recycling sector is growing considerably in Europe due to increasingly strict regulations aimed at inciting the consumption of low energy light bulbs and their end-of-life management. Chemical risks were assessed in fluorescent lamp recycling facilities by field measurement surveys in France, highlighting that occupational exposure and pollutant levels in the working environment were correlated with the main recycling steps and processes.The mean levels of worker exposure are 4.4 mg/m³, 15.4 μg/m³, 14.0 μg/m³, 247.6 μg/m³, respectively, for total inhalable dust, mercury, lead and yttrium. The mean levels of airborne pollutants are 3.1 mg/m³, 9.0 μg/m³, 9.0 μg/m³, 219.2 μg/m³, respectively, for total inhalable dust, mercury, lead and yttrium. The ranges are very wide. Surface samples from employees’ skin and granulometric analysis were also carried out. The overview shows that all the stages and processes involved in lamp recycling are concerned by the risk of hazardous substances penetrating into the bodies of employees, although exposure of the latter varies depending on the processes and tasks they perform. The conclusion of this study strongly recommends the development of a new generation of processes in parallel with more information sharing and regulatory measures.