联合GF-6和Sentinel-2红边波段的森林地上生物量反演

光谱反射率能反映地物差异,是森林地上生物量（Aboveground Biomass,AGB）遥感反演的理论基础。红边波段处于近红外与红光波段交界处快速变化的区域,能对植被冠层结构和叶绿素含量的微小变化做出快速反应,对植被生长状况较敏感。研究以GF-6和Sentinel-2多光谱影像作为数据源,结合野外调查AGB数据,构建落叶松和樟子松AGB线性和非线性估测模型,通过比较模型精度选择最优模型进行森林AGB反演和空间分布制图。结果表明：GF-6和Sentinel-2影像红边波段反射率与落叶松、樟子松AGB均呈显著相关（P<0.05）,红边波段对AGB估测较敏感。多变量估测模型整体估测效果优于单变量模型,所有模型中多元线性回归模型取得了最优的决定系数（落叶松R²=0.66,樟子松R²=0.65）和最低的均方根误差（落叶松RMSE=31.45 t/hm²,樟子松RMSE=54.77 t/hm²）。相比单个数据源,联合GF-6和Sentinel-2影像构建的多元线性回归模型估测效果得到了显著提升,模型RMSE对于落叶松和樟子松AGB估测分别最大降低了22.9%和11.2%。增加红边波段进行AGB估测能显著提高模型估测精度,三组数据源分别加入红边波段信息后进行建模,模型RMSE得到了显著降低。GF-6拥有800 km观测幅宽和高效的重访周期,可以快速地提供大尺度时间序列数据,在森林地上生物量反演和动态监测方面有着很大潜力。     

Allometric equations to predict the total aboveground biomass of tree species in a planted forest in Egypt

The aim of this study is to estimate the total above‐ground biomass (TAGB), stem height (H), diameter at breast height (dbh) and basal area of five tree species (ages 7‐8 years) irrigated by municipal sewage water in the Egyptian‐Chinese friendship forest, Sadat City, Egypt. From the biomass data that obtained through destructive sampling, models for predicting aboveground biomass were developed. The highest values for stem density and height were estimated for Eucalyptus citriodora, while the lowest value for density was obtained for Dalbergia sissoo and stem height for Khaya senegalensis. The highest values for basal area and dbh were obtained for Casuarina spp., while the lowest values were recorded for Dalbergia sissoo. Eucalyptus camaldulensis had the highest stand stem biomass and TAGB (55.5, 83.9 t DW ha^‐1, respectively). In addition, Casuarina spp. had the highest leafy branches biomass (32.5 t DW ha^‐1) while Dalbergia sissoo had the lowest values for all tree components. All the generated allometric equations had high correlation coefficients at high probability levels. Moreover, the results revealed that not only the dbh data can be used as independent variable for biomass determination, but also stem height and size index are recommended for biomass estimation (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Standing crop and aboveground biomass partitioning of a dwarf mangrove forest in Taylor River Slough,Florida

The structure and standing crop biomass of a dwarf mangrove forest, located in the salinity transition zone ofTaylor River Slough in the Everglades National Park, were studied. Although the four mangrove species reported for Florida occurred at the study site, dwarf Rhizophora mangle trees dominated the forest. The structural characteristics of the mangrove forest were relatively simple: tree height varied from 0.9 to 1.2 meters, and tree density ranged from 7062 to 23 778 stems ha^–1. An allometric relationship was developed to estimate leaf, branch, prop root, and total aboveground biomass of dwarf Rhizophora mangle trees. Total aboveground biomass and their components were best estimated as a power function of the crown area times number of prop roots as an independent variable (Y = B × X^–0.5083). The allometric equation for each tree component was highly significant (p<0.0001), with all r² values greater than 0.90. The allometric relationship was used to estimate total aboveground biomass that ranged from 7.9 to 23.2 ton ha^–1. Rhizophora mangle contributed 85% of total standing crop biomass. Conocarpus erectus, Laguncularia racemosa, and Avicennia germinans contributed the remaining biomass. Average aboveground biomass allocation was 69% for prop roots, 25% for stem and branches, and 6% for leaves. This aboveground biomass partitioning pattern, which gives a major role to prop roots that have the potential to produce an extensive root system, may be an important biological strategy in response to low phosphorus availability and relatively reduced soils that characterize mangrove forests in South Florida.     

Estimation of aboveground biomass and net biomass increment in a cool temperate forest on a landscape scale

I clarified aboveground biomass (AGB), net biomass increment (NBI) and its spatial heterogeneity in a cool temperate forest on a landscape scale (>2,200 ha). The relationships among AGB, NBI, and the size frequency distribution of trees of each stand were examined by combining an analysis of vegetation using aerial photographs, and data from 146 inventory plots (28.8 ha in total). This area included natural broad-leaved stands, harvested broad-leaved stands, and artificial conifer plantations. A –3/2 power distribution density function was applied to the individual mass frequency distribution of each plot. Estimated AGB in carbon (C) equivalent was 480–5,615 g C m^–2 (3,130 g C m^–2 on average), and NBI was –98 to 436 g C m^–2 year^–1 (83.0 g C m^–2 year^–1 on average). NBI had a single significant relationship with the reciprocal of theoretical maximum individual mass, while NBI was not significantly related to AGB. My results showed that, on a landscape scale, AGB and NBI strongly depend on the size structure of forest stands.     

Loss of aboveground forest biomass and landscape biomass variability in Missouri,US

Disturbance regimes and forests have changed over time in the eastern United States. We examined effects of historical disturbance (circa 1813 to 1850) compared to current disturbance (circa 2004 to 2008) on aboveground, live tree biomass (for trees with diameters ≥13 cm) and landscape variation of biomass in forests of the Ozarks and Plains landscapes in Missouri, USA. We simulated 10,000 one-hectare plots using random diameters generated from parameters of diameter distributions limited to diameters ≥13 cm and random densities generated from density estimates. Area-weighted mean biomass density (Mg/ha) for historical forests averaged 116 Mg/ha, ranging from 54 Mg/ha to 357 Mg/ha by small scale ecological subsections within Missouri landscapes. Area-weighted mean biomass density for current forests averaged 82 Mg/ha, ranging from 66 Mg/ha to 144 Mg/ha by ecological subsection for currently forested land. Biomass density of current forest was greater than historical biomass density for only 2 of 23 ecological subsections. Current carbon sequestration of 292 TgC on 7 million ha of forested land is less than half of the estimated historical total carbon sequestration of 693 TgC on 12 million ha. Cumulative tree cutting disturbances over time have produced forests that have less aboveground tree biomass and are uniform in biomass compared to estimates of historical biomass, which varied across Missouri landscapes. With continued relatively low rates of forest disturbance, current biomass per ha will likely increase to historical levels as the most competitive trees become larger in size and mean number of trees per ha decreases due to competition and self-thinning. Restoration of large diameter structure and forested extent of upland woodlands and floodplain forests could fulfill multiple conservation objectives, including carbon sequestration.     

Sulfur speciation and distribution in soils and aboveground biomass of a boreal coniferous forest

Major sulfur pools are quantified in soils and aboveground biomass of a coniferous boreal forest. Total ecosystem S averages 1395 kg·ha⁻¹ of which 98% is found in the soil, with 89% being in the mineral horizons. Organic S dominates in soil, tree parts (trunks, branches + foliage, roots) and litterfall, ranging from 77 to 99% of total S concentration. Carbon-bonded S, ester sulfate and SO₄-S in soil profiles range respectively from 51–68%, 29–37% and 1–14% of total S concentrations and account respectively for 57, 33 and 10% of total S on an areal basis. Adsorbed SO₄ accounts for 82% of total SO₄, and can be predicted from Al bound to organic matter, amorphous Al and pH (r^{2 = 0.81}). There is a strong relationship between % carbon and carbon-bonded S in 4 of the 5 soil horizons studied which represent over 95% of the total soil organic matter, whereas ester sulfate is related to % carbon in 3 soil horizons representing only 37% of the soil organic matter. An analysis of atmospheric S loading and S data for 10 forested sites in Europe and North America suggests that the size of the organic S pool in forested systems is independent of atmospheric loading.     

星载大光斑LiDAR与HJ-1A高光谱数据联合估测区域森林地上生物量

以吉林省汪清林业局经营区为研究区,利用HJ-1A/HSI高光谱数据和ICESat-GLAS波形数据,估测区域森林地上生物量。从平滑后的GLAS波形数据中提取波形长度W和地形坡度参数TS,建立GLAS森林最大树高估测模型;从GLAS波形数据中提取能量参数I(植被回波能量Ev和回波总能量E之比),建立GLAS森林郁闭度估测模型;利用GLAS估测的森林最大树高和森林郁闭度联合建立森林地上生物量模型。由于GLAS呈离散条带状分布,无法实现区域估测,因此研究将GLAS波形数据与HJ-1A/HSI高光谱数据联合,基于支持向量回归机算法实现森林地上生物量区域估测,得到研究区森林地上生物量分布图。研究结果显示,基于W和TS建立的GLAS森林最大树高估测模型的adj.R~2=0.78,RMSE=2.51m,模型验证的adj.R~2=0.85,RMSE=1.67m。地形坡度参数TS能够有效的降低地形坡度的影响;当林下植被高度为2m时,得到的基于参数I建立的GLAS森林郁闭度估测模型效果最好,模型的adj.R~2=0.64,RMSE=0.13,模型验证的adj.R~2=0.65,RMSE=0.12。利用森林最大树高和森林郁闭度建立的森林地上生物量模型的adj.R~2=0.62,RMSE=10.88 t/hm~2,模型验证的adj.R~2=0.60,RMSE=11.52 t/hm~2。基于支持向量回归机算法,利用HJ-1A/HSI和GLAS数据建立的森林地上生物量SVR模型,生成了森林地上生物量分布图,利用野外数据对得到的分布图进行验证,验证结果显示森林地上生物量估测值与实测值存在很强的线性关系(adj.R~2=0.62,RMSE=11.11 t/hm~2),能够满足林业应用的需要。因此联合ICESat-GLAS波形数据与HJ-1A高光谱数据,能够提高区域森林地上生物量的估测精度。     

A new disease-specific machine learning approach for the prediction of cancer-causing missense variants

High-throughput genotyping and sequencing techniques are rapidly and inexpensively providing large amounts of human genetic variation data. Single Nucleotide Polymorphisms (SNPs) are an important source of human genome variability and have been implicated in several human diseases, including cancer. Amino acid mutations resulting from non-synonymous SNPs in coding regions may generate protein functional changes that affect cell proliferation. In this study, we developed a machine learning approach to predict cancer-causing missense variants. We present a Support Vector Machine (SVM) classifier trained on a set of 3163 cancer-causing variants and an equal number of neutral polymorphisms. The method achieve 93% overall accuracy, a correlation coefficient of 0.86, and area under ROC curve of 0.98. When compared with other previously developed algorithms such as SIFT and CHASM our method results in higher prediction accuracy and correlation coefficient in identifying cancer-causing variants.     

Palm live aboveground biomass in the riparian zones of a forest in Central Amazonia

Aboveground biomass estimates in the Amazon region remain uncertain, partly due to extrapolations based mainly on samples collected in upland terrains of terra-firme forests. Most biomass estimates were focused on dicotyledonous trees or included other plant groups as a category of trees. Palms dominate areas that represent 20% of the Brazilian Amazon. However, their contribution to biomass estimates and the variation within riparian zones remain poorly documented. We estimated the biomass of palms larger than 1–cm diameter at breast height (1.3 m aboveground) in riparian plots (n = 40); investigated the potential bias caused by the use of dicotyledonous- or family- rather than species-level equations for biomass estimation; compared palm biomass between riparian and non-riparian plots (n = 72); and evaluated the effects of soil, topography, and stream characteristics in riparian plots on palm biomass. Mean palm biomass in riparian zones estimated with species-level equations (27.50 ± 12.94 Mg/ha, range: 3.32–63.27 Mg/ha) was three times greater than biomass estimated with a family-level equation (9.04 ± 4.29 Mg/ha, range: 1.51–21.25 Mg/ha) and was greater than mean biomass estimated with a pantropical equation (20.46 ± 9.29 Mg/ha, range: 3.67–47.99 Mg/ha). Mean palm biomass in riparian zones was four times greater than in non-riparian zones. Palm biomass was high in flatter areas with poorly drained soils, but lower around streams with higher discharge. Inclusion of palms can contribute to reducing the uncertainties in biomass estimates in Amazonian forests. Recognition of the importance of riparian zones may improve conservation policies. Abstract in Portuguese is available with online material.     

Estimating fine-root biomass and production of boreal and cool temperate forests using aboveground measurements: A new approach

Information of fine-root biomass and production is critical for quantifying the productivity and carbon cycle of forest ecosystems, and yet our ability to obtain this information especially at a large spatial scale (e.g., regional to global) is extremely limited. Several studies attempted to relate fine-root biomass and production with various aboveground variables that can be measured more easily so that fine-root biomass and production could be estimated at a large spatial scale, but found the correlations were generally weak or non-existed at the stand level. In this study, we tested a new approach: instead of using the conventional way of analysing fine-root biomass at the stand level, we analysed fine-root data at the tree level. Fine-root biomass of overstory trees in stand was first separated from that of understory and standardized to a common fine-root definition of < 2 mm or < 5 mm diameter. Afterwards, we calculated fine-root biomass per tree for a representative tree size of mean basal area for each stand. Statistically significant correlations between the fine-root biomass per tree and the diameter at the ground surface were found for all four boreal and cool temperate spruce, pine, fir and broadleaf forest types, and so allometric equations were developed for each group using a total of n = 212 measurements. The stand-level fine-root biomass of trees estimated using the allometric equations agrees well with the measurements, with r ² values of 0.64 and 0.57 (n = 171), respectively, for fine-roots < 2 mmand < 5 mm diameter. This study further estimated fine-root production as the product of fine-root turnover rate and fine-root biomass, and determined the turnover rate as a function of fine-root biomass, stand age, and mean annual temperature. The estimates of tree fine-root production agree well with reported values, with r ² value of 0.53 for < 2 mm and 0.54 for < 5 mm diameter (n = 162) at the stand level.     

A field based statistical approach for validating a remotely sensed mangrove forest classification scheme

Amongst the most threatened ecosystems on Earth, mangrove forests are also one of the more difficult to work in due to their growth in mud and open water coastal zones and their dense tangled stems, branches and prop roots. Consequently, there has been an impetus to employ remotely sensed imagery as a means for rapid inventory of these coastal wetlands. To date, the majority of mangrove maps derived from satellite imagery utilize a simple mangrove classification scheme which does not distinguish mangrove species and may not be useful for conservation and management purposes. Although more elaborate satellite based mangrove classification schemes are being developed, given their enhanced complexity they deserve additional justification for end users. The purpose of this study was to statistically examine the appropriateness of one such classification scheme based on an inventory of field data. In January of 2007 and May of 2008, 61 field sample plots were selected in a stratified random fashion based on a previous classification of a degraded mangrove forest of the Isla La Palma (Sinaloa, Mexico) using Landsat TM5 data. Unlike other previous Landsat TM based classifications of this region, which simply identified the mangrove forests as one class, the mangroves were classified (i.e. mapped) according to four conditions; healthy tall, healthy dwarf, poor condition, and dead mangroves. Within each sample plot, all mangroves of diameter of breast height (dbh) greater than 2.5 cm were identified and their height, condition and dbh recorded. An estimated Leaf Area Index (LAI) value also was obtained for each sample and the shortest distance from the center of each sample plot to open flowing water was determined using a geographic information system (GIS) overlay procedure. These data were then used to calculate mean values for the four classes as well as to determine stem densities, basal areas, and the Shannon–Wiener diversity index. In order to assess the appropriateness of this mangrove classification scheme a discriminant analysis approach was then applied to these field data. The results indicate this forest has undergone severe degradation, with decreasing mean tree heights, mean dbh and species diversity. In regards to the discriminant analysis procedure, further classification of these field plots and cross-validation based on these significant variables provided high classification accuracy thus validating the appropriateness of the satellite based image classification scheme. Moreover, the discriminant analysis indicated that the estimated LAI, mean height, and mean dbh are significant in the separation of the classification of mangrove forest condition along these field sample plots.     

Influences of vegetation,model, and data parameters on forest aboveground biomass assessment using an area-based approach

Although aboveground biomass (AGB) estimation using area-based approaches (ABAs) and its application to forestry have been actively researched through three decades, this technology has been little operationalized in the Central European forest sector. That means specific recommendations are needed in order to apply ABA for forest biomass modelling in this region. The present study was directed to filling such gaps while examining the effect of input ABA parameters on AGB model quality in conditions of mixed mountainous forests in Central Europe. Specific objectives were to assess whether the strength of the AGB model can be impacted by 1) canopy conditions (leaf-on and leaf-off), 2) airborne LiDAR point density (2.5, 5.0, 7.5, 10.0 points/m²), 3) field methods to estimate AGB (with regeneration components or without), and 4) machine learning methods (AdaBoost, Random decision forest, multilayer neural network, and Bayesian ridge regression). The results show that canopy conditions and airborne LiDAR point densities did not affect the strength of the AGB model, but that model's strength was affected by the vegetation regeneration component in the field biomass reference and by the machine learning method tested for modelling. AdaBoost and random decision forest were the most successful methods. To evaluate the quality of an AGB model it is recommended to combine several individual evaluation functions into the model score. The study highlights several recommendations to follow when estimating AGB from ALS using an ABA in Central European forests.     

Patterns of mangrove forest disturbance and biomass removal due to small-scale harvesting in southwestern Madagascar

Wetlands Ecology and Management - Informal small-scale mangrove wood harvesting has received limited attention, though it is a widespread threat to mangroves in many parts of the tropics. We...     

Fusion of airborne LiDAR data and hyperspectral imagery for aboveground and belowground forest biomass estimation

Vegetation biomass is a key biophysical parameter for many ecological and environmental models. The accurate estimation of biomass is essential for improving the accuracy and applicability of these models. Light Detection and Ranging (LiDAR) data have been extensively used to estimate forest biomass. Recently, there has been an increasing interest in fusing LiDAR with other data sources for directly measuring or estimating vegetation characteristics. In this study, the potential of fused LiDAR and hyperspectral data for biomass estimation was tested in the middle Heihe River Basin, northwest China. A series of LiDAR and hyperspectral metrics were calculated to obtain the optimal biomass estimation model. To assess the prediction ability of the fused data, single and fused LiDAR and hyperspectral metrics were regressed against field-observed belowground biomass (BGB), aboveground biomass (AGB) and total forest biomass (TB). The partial least squares (PLS) regression method was used to reduce the multicollinearity problem associated with the input metrics. It was found that the estimation accuracy of forest biomass was affected by LiDAR plot size, and the optimal plot size in this study had a radius of 22 m. The results showed that LiDAR data alone could estimate biomass with a relative high accuracy, and hyperspectral data had lower prediction ability for forest biomass estimation than LiDAR data. The best estimation model was using a fusion of LiDAR and hyperspectral metrics (R² = 0.785, 0.893 and 0.882 for BGB, AGB and TB, respectively, with p < 0.0001). Compared with LiDAR metrics alone, the fused LiDAR and hyperspectral data improved R² by 5.8%, 2.2% and 2.6%, decreased AIC value by 1.9%, 1.1% and 1.2%, and reduced RMSE by 8.6%, 7.9% and 8.3% for BGB, AGB and TB, respectively. These results demonstrated that biomass accuracies could be improved by the use of fused LiDAR and hyperspectral data, although the improvement was slight when compared with LiDAR data alone. This slight improvement could be attributed to the complementary information contained in LiDAR and hyperspectral data. In conclusion, fusion of LiDAR and other remotely sensed data has great potential for improving biomass estimation accuracy.     

A machine learning approach to identify hydrogenosomal proteins in Trichomonas vaginalis

The protozoan parasite Trichomonas vaginalis is the causative agent of trichomoniasis, the most widespread nonviral sexually transmitted disease in humans. It possesses hydrogenosomes-anaerobic mitochondria that generate H(2), CO(2), and acetate from pyruvate while converting ADP to ATP via substrate-level phosphorylation. T. vaginalis hydrogenosomes lack a genome and translation machinery; hence, they import all their proteins from the cytosol. To date, however, only 30 imported proteins have been shown to localize to the organelle. A total of 226 nuclear-encoded proteins inferred from the genome sequence harbor a characteristic short N-terminal presequence, reminiscent of mitochondrial targeting peptides, which is thought to mediate hydrogenosomal targeting. Recent studies suggest, however, that the presequences might be less important than previously thought. We sought to identify new hydrogenosomal proteins within the 59,672 annotated open reading frames (ORFs) of T. vaginalis, independent of the N-terminal targeting signal, using a machine learning approach. Our training set included 57 gene and protein features determined for all 30 known hydrogenosomal proteins and 576 nonhydrogenosomal proteins. Several classifiers were trained on this set to yield an import score for all proteins encoded by T. vaginalis ORFs, predicting the likelihood of hydrogenosomal localization. The machine learning results were tested through immunofluorescence assay and immunodetection in isolated cell fractions of 14 protein predictions using hemagglutinin constructs expressed under the homologous SCSα promoter in transiently transformed T. vaginalis cells. Localization of 6 of the 10 top predicted hydrogenosome-localized proteins was confirmed, and two of these were found to lack an obvious N-terminal targeting signal.     

The assessment of mangrove biomass and carbon in West Africa: a spatially explicit analytical framework

Mangrove forests are highly productive and have large carbon sinks while also providing numerous goods and ecosystem services. However, effective management and conservation of the mangrove forests are often dependent on spatially explicit assessments of the resource. Given the remote and highly dispersed nature of mangroves, estimation of biomass and carbon in mangroves through routine field-based inventories represents a challenging task which is impractical for large-scale planning and assessment. Alternative approaches based on geospatial technologies are needed to support this estimation in large areas. However, spatial data processing and analysis approaches used in this estimation of mangrove biomass and carbon have not been adequately investigated. In this study, we present a spatially explicit analytical framework that integrate remotely sensed data and spatial analyses approaches to support the estimation of mangrove biomass and carbon stock and their spatial patterns in West Africa. Forest canopy height derived from SRTM and ICESat/GLAS data was used to estimate mangrove biomass and carbon in nine West African countries. We developed a geospatial software toolkit that implemented the proposed framework. The spatial analysis framework and software toolkit provide solid support for the estimation and relative comparisons of mangrove-related metrics. While the mean canopy height of mangroves in our study area is 10.2 m, the total biomass and carbon were estimated as 272.56 and 136.28 Tg. Nigeria has the highest total mangrove biomass and carbon in the nine countries, but Cameroon is the country with the largest mean biomass and carbon density. The resulting spatially explicit distributions of mangrove biomass and carbon hold great potential in guiding the strategic planning of large-scale field-based assessment of mangrove forests. This study demonstrates the utility of online geospatial data and spatial analysis as a feasible solution for estimating the distribution of mangrove biomass and carbon at larger or smaller scales.     

Functional identity of overstorey tree height and understorey conservative traits drive aboveground biomass in a subtropical forest

The niche complementarity hypothesis has received empirical support but species differ in functional strategies for their contribution to ecosystem function, as predicted by the mass ratio hypothesis. Our understanding of how functional identity of conservative and acquisitive strategies of trees predicts aboveground biomass across forest strata (i.e. overstorey and understorey) remains unclear. Aboveground biomass, community-weighted mean (CWM − functional identity) of trait values (6 leaf and 2 stem traits), and soil physicochemical properties were estimated for 125 plots in a 5-ha subtropical forest in Eastern China. We used multiple linear regressions models to relate aboveground biomass to CWM indices at overstorey and understorey strata separately, and whole-community level. We finally employed the structural equation model to test for the effects of overstorey on understorey strata, in addition to the effects of soil physicochemical properties. Forest strata optimal models showed that overstorey strata had high aboveground biomass when they are dominated by functional identity of tree height, whereas high aboveground biomass in understorey strata was driven by functional identity of dense-wooded conservative strategy. Whole-community optimal model showed that communities dominated by functional identity of leaf dry matter content and mean leaf area had high aboveground biomass. Aboveground biomass was negatively related to soil nutrients across forest strata and whole-community level. The structural equation model showed that CWM of overstorey tree height did not affect understorey functional identity and aboveground biomass, when soil physicochemical properties were accounted. Soil nutrients had positive effect on functional identity of overstorey tree height whereas negative effect on functional identity of understorey dense-wooded strategy. This study highlights the fundamental roles of forest strata where overstorey and understorey strata contribute to their corresponding aboveground biomass with contrasting functional strategies across a range of soil nutrients. High aboveground biomass was potentially driven by functional identity of tree height through making use of plentiful soil nutrients at overstorey strata, whereas by conservative strategy at understorey strata through enduring nutrient-poor soils. To better understand the roles of functional identity of conservative and acquisitive strategies in driving ecosystem functions, it is worth to analyse forest strata separately.     

Using machine learning to predict the effects and consequences of mutations in proteins

Machine and deep learning approaches can leverage the increasingly available massive datasets of protein sequences, structures, and mutational effects to predict variants with improved fitness. Many different approaches are being developed, but systematic benchmarking studies indicate that even though the specifics of the machine learning algorithms matter, the more important constraint comes from the data availability and quality utilized during training. In cases where little experimental data are available, unsupervised and self-supervised pre-training with generic protein datasets can still perform well after subsequent refinement via hybrid or transfer learning approaches. Overall, recent progress in this field has been staggering, and machine learning approaches will likely play a major role in future breakthroughs in protein biochemistry and engineering.     

Tree diversity,composition, forest structure and aboveground biomass dynamics after single and repeated fire in a Bornean rain forest

Forest fires remain a devastating phenomenon in the tropics that not only affect forest structure and biodiversity, but also contribute significantly to atmospheric CO2. Fire used to be extremely rare in tropical forests, leaving ample time for forests to regenerate to pre-fire conditions. In recent decades, however, tropical forest fires occur more frequently and at larger spatial scales than they used to. We studied forest structure, tree species diversity, tree species composition, and aboveground biomass during the first 7 years since fire in unburned, once burned and twice burned forest of eastern Borneo to determine the rate of recovery of these forests. We paid special attention to changes in the tree species composition during burned forest regeneration because we expect the long-term recovery of aboveground biomass and ecosystem functions in burned forests to largely depend on the successful regeneration of the pre-fire, heavy-wood, species composition. We found that forest structure (canopy openness, leaf area index, herb cover, and stem density) is strongly affected by fire but shows quick recovery. However, species composition shows no or limited recovery and aboveground biomass, which is greatly reduced by fire, continues to be low or decline up to 7 years after fire. Consequently, large amounts of the C released to the atmosphere by fire will not be recaptured by the burned forest ecosystem in the near future. We also observed that repeated fire, with an inter-fire interval of 15 years, does not necessarily lead to a huge deterioration in the regeneration potential of tropical forest. We conclude that burned forests are valuable and should be conserved and that long-term monitoring programs in secondary forests are necessary to determine their recovery rates, especially in relation to aboveground biomass accumulation.