Cornelius Senf

Mapping Rubber Plantations and Natural Forests in Xishuangbanna (Southwest China) Using Multi-Spectral Phenological Metrics from MODIS Time Series

We developed and evaluated a new approach for mapping rubber plantations and natural forests in one of Southeast Asia’s biodiversity hot spots, Xishuangbanna in China. We used a one-year annual time series of Moderate Resolution Imaging Spectroradiometer (MODIS), Enhanced Vegetation Index (EVI) and short-wave infrared (SWIR) reflectance data to develop phenological metrics. These phenological metrics were used to classify rubber plantations and forests with the Random Forest classification algorithm. We evaluated which key phenological characteristics were important to discriminate rubber plantations and natural forests by estimating the influence of each metric on the classification accuracy. As a benchmark, we compared the best classification with a classification based on the full, fitted time series data. Overall classification accuracies derived from EVI and SWIR time series alone were 64.4% and 67.9%, respectively. Combining the phenological metrics from EVI and SWIR time series improved the accuracy to 73.5%. Using the full, smoothed time series data instead of metrics derived from the time series improved the overall accuracy only slightly (1.3%), indicating that the phenological metrics were sufficient to explain the seasonal changes captured by the MODIS time series. The results demonstrate a promising utility of phenological metrics for mapping and monitoring rubber expansion with MODIS.

Estimating Fractional Shrub Cover Using Simulated EnMAP Data: A Comparison of Three Machine Learning Regression Techniques

Anthropogenic interventions in natural and semi-natural ecosystems often lead to substantial changes in their functioning and may ultimately threaten ecosystem service provision. It is, therefore, necessary to monitor these changes in order to understand their impacts and to support management decisions that help ensuring sustainability. Remote sensing has proven to be a valuable tool for these purposes, and especially hyperspectral sensors are expected to provide valuable data for quantitative characterization of land change processes. In this study, simulated EnMAP data were used for mapping shrub cover fractions along a gradient of shrub encroachment, in a study region in southern Portugal. We compared three machine learning regression techniques: Support Vector Regression (SVR); Random Forest Regression (RF); and Partial Least Squares Regression (PLSR). Additionally, we compared the influence of training sample size on the prediction performance. All techniques showed reasonably good results when trained with large samples, while SVR always outperformed the other algorithms. The best model was applied to produce a fractional shrub cover map for the whole study area. The predicted patterns revealed a gradient of shrub cover between regions affected by special agricultural management schemes for nature protection and areas without land use incentives. Our results highlight the value of EnMAP data in combination with machine learning regression techniques for monitoring gradual land change processes.

Remote sensing of forest insect disturbances: Current state and future directions

Insect disturbance are important agents of change in forest ecosystems around the globe, yet their spatial and temporal distribution and dynamics are not well understood. Remote sensing has gained much attention in mapping and understanding insect outbreak dynamics. Consequently, we here review the current literature on the remote sensing of insect disturbances. We suggest to group studies into three insect types: bark beetles, broadleaved defoliators, and coniferous defoliators. By so doing, we systematically compare the sensors and methods used for mapping insect disturbances within and across insect types. Results suggest that there are substantial differences between methods used for mapping bark beetles and defoliators, and between methods used for mapping broadleaved and coniferous defoliators. Following from this, we highlight approaches that are particularly suited for each insect type. Finally, we conclude by highlighting future research directions for remote sensing of insect disturbances. In particular, we suggest to: 1) Separate insect disturbances from other agents; 2) Extend the spatial and temporal domain of analysis; 3) Make use of dense time series; 4) Operationalize near-real time monitoring of insect disturbances; 5) Identify insect disturbances in the context of coupled human-natural systems; and 6) Improve reference data for assessing insect disturbances. Since the remote sensing of insect disturbances has gained much interest beyond the remote sensing community recently, the future developments identified here will help integrating remote sensing products into operational forest management. Furthermore, an improved spatiotemporal quantification of insect disturbances will support an inclusion of these processes into regional to global ecosystem models.

Using Landsat time series for characterizing forest disturbance dynamics in the coupled human and natural systems of Central Europe

Remote sensing is a key information source for improving the spatiotemporal understanding of forest ecosystem dynamics. Yet, the mapping and attribution of forest change remains challenging, particularly in areas where a number of interacting disturbance agents simultaneously affect forest development. The forest ecosystems of Central Europe are coupled human and natural systems, with natural and human disturbances affecting forests both individually and in combination. To better understand the complex forest disturbance dynamics in such systems, we utilize 32-year Landsat time series to map forest disturbances in five sites across Austria, the Czech Republic, Germany, Poland, and Slovakia. All sites consisted of a National Park and the surrounding forests, reflecting three management zones of different levels of human influence (managed, protected, strictly protected). This allowed for a comparison of spectral, temporal, and spatial disturbance patterns across a gradient from natural to coupled human and natural disturbances. Disturbance maps achieved overall accuracies ranging from 81% to 93%. Disturbance patches were generally small, with 95% of the disturbances being smaller than 10 ha. Disturbance rates ranged from 0.29% yr-1 to 0.95% yr-1, and differed substantially among management zones and study sites. Natural disturbances in strictly protected areas were longer in duration (median of 8 years) and slightly less variable in magnitude compared to human-dominated disturbances in managed forests (median duration of 1 year). However, temporal dynamics between natural and human-dominated disturbances showed strong synchrony, suggesting that disturbance peaks are driven by natural events affecting managed and unmanaged areas simultaneously. Our study demonstrates the potential of remote sensing for mapping forest disturbances in coupled human and natural systems, such as the forests of Central Europe. Yet, we also highlight the complexity of such systems in terms of agent attribution, as many natural disturbances are modified by management responding to them outside protected areas.

Natural disturbances are spatially diverse but temporally synchronized across temperate forest landscapes in Europe

Natural disturbance regimes are changing substantially in forests around the globe. However, large-scale disturbance change is modulated by a considerable spatiotemporal variation within biomes. This variation remains incompletely understood particularly in the temperate forests of Europe, for which consistent large-scale disturbance information is lacking. Here, our aim was to quantify the spatiotemporal patterns of forest disturbances across temperate forest landscapes in Europe using remote sensing data and determine their underlying drivers. Specifically, we tested two hypotheses: (1) Topography determines the spatial patterns of disturbance, and (2) climatic extremes synchronize natural disturbances across the biome. We used novel Landsat-based maps of forest disturbances 1986-2016 in combination with landscape analysis to compare spatial disturbance patterns across five unmanaged forest landscapes with varying topographic complexity. Furthermore, we analyzed annual estimates of disturbances for synchronies and tested the influence of climatic extremes on temporal disturbance patterns. Spatial variation in disturbance patterns was substantial across temperate forest landscapes. With increasing topographic complexity, natural disturbance patches were smaller, more complex in shape, more dispersed, and affected a smaller portion of the landscape. Temporal disturbance patterns, however, were strongly synchronized across all landscapes, with three distinct waves of high disturbance activity between 1986 and 2016. All three waves followed years of pronounced drought and high peak wind speeds. Natural disturbances in temperate forest landscapes of Europe are thus spatially diverse but temporally synchronized. We conclude that the ecological effect of natural disturbances (i.e., whether they are homogenizing a landscape or increasing its heterogeneity) is strongly determined by the topographic template. Furthermore, as the strong biome-wide synchronization of disturbances was closely linked to climatic extremes, large-scale disturbance episodes are likely in Europes temperate forests under climate changes.

Using Landsat to Assess the Relationship Between Spatiotemporal Patterns of Western Spruce Budworm Outbreaks and Regional-Scale Weather Variability

Abstract. Western spruce budworm is a defoliator of coniferous forests of North America. Past research has shown that climate is a principal driver of budworm outbreaks, though the underlying relationships are not well understood. We utilized Landsat time series to investigate the relationships between spatiotemporal patterns of budworm outbreaks and weather variability. Landsat-based maps of budworm infestations from (1995–2013) were produced and used to describe patterns of the most recent budworm outbreak in British Columbia. Superposed epoch and regression analysis were used to explore relationships between outbreak patterns and annual temperature and precipitation anomalies. We found that western spruce budworm outbreaks were preceded by autumn precipitation deficits that were 12% lower than the long-term average, and co-occurred with summer precipitation up to 20% below the long-term average. Spring temperatures shortly before outbreaks were 0.7°C lower than normal, whereas winter temperatures during outbreaks were above average by 0.5°C. We concluded that the most recent outbreak of western spruce budworm in British Columbia co-occurred with distinct weather patterns that probably facilitated budworm population development and synchronized host-herbivore phenology. Although Landsat has been used to map insect disturbances, our study demonstrated its usefulness for understanding landscape- to regional-scale drivers of budworm outbreaks.

Agriculture rivals biomes in predicting global species richness

Species-area relationships (SARs) provide an avenue to model patterns of species richness and have recently been shown to vary substantially across regions of different climate, vegetation, and land cover. Given that a large proportion of the globe has been converted to agriculture, and considering the large variety in agricultural management practices, a key question is whether global SARs vary across gradients of agricultural intensity. We developed SARs for mammals that account for geographic variation in biomes, land cover and a range of land-use intensity indicators representing inputs (e.g. fertilizer, irrigation), outputs (e.g. yields) and system-level measures of intensity (e.g. human appropriation of net primary productivity - HANPP). We systematically compared the resulting SARs in terms of their predictive ability. Our global SAR with a universal slope was significantly improved by the inclusion of any one of the three variable types: biomes, land cover, and land-use intensity. The latter, in the form of human appropriation of net primary productivity (HANPP), performed as well as biomes and land-cover in predicting species richness. Other land-use intensity indicators had a lower predictive ability. Our main finding that land-use intensity performs as well as biomes and land cover in predicting species richness emphasizes that human factors are on a par with environmental factors in predicting global patterns of biodiversity. While our broad-scale study cannot establish causality, human activity is known to drive species richness at a local scale, and our findings suggest that this may hold true at a global scale. The ability of land-use intensity to explain variation in SARs at a global scale had not previously been assessed. Our study suggests that the inclusion of land-use intensity in SAR models allows us to better predict and understand species richness patterns. This article is protected by copyright. All rights reserved.

Using MODIS time series and random forests classification for mapping land use in South-East Asia

Broad scale and continuous land-use/cover mapping is important for research in the context of global and climate change. We have therefore developed a method based on MODIS time-series and Random Forest classification to map forested, non-forested and plantation areas in South-East Asia. Our approach is optimized for regions with frequent cloud cover and scarce reference data. Results show that the method performs acceptable in terms of accuracy under given conditions. Furthermore, various validation strategies were compared with the result that pixel-by-pixel validation procedures with manual assignment must be handled carefully when applied to large scale land-use/cover maps.

Comparing Support Vector Regression and Statistical Linear Regression for Predicting Poverty Incidence in Vietnam

Urban and rural poverty are key issues of the Millennium Development Goals and much research is done on how to reduce poverty sustainable and long-ranging. However, smallscalepovertymaps at full spatial and temporalcoverage are fundamentallynecessary but rare. Some small scale poverty mapping methods have been developed in past years, but these methods often rely on data which has to be collected in resource intensive field work. We therefore compare two statistical data mining tools, Support Vector Regression and Linear Regression, to scale Vietnamese poverty data from a coarser training to smaller scaled testing set. The Support Vector Regression performedworse than the Linear Regression model with feature subset. However, the Support Vector Regression model showed a more systematic error which might be corrected more easily than the error of the Linear Regression approach. Furthermore, both models showed dependency on spatial effects. Hence, integration of spatial information might increase the success of future models and turn data mining approaches into valuable tools for poverty mapping on small scales.

Patterns and drivers of recent disturbances across the temperate forest biome

Increasing evidence indicates that forest disturbances are changing in response to global change, yet local variability in disturbance remains high. We quantified this considerable variability and analyzed whether recent disturbance episodes around the globe were consistently driven by climate, and if human influence modulates patterns of forest disturbance. We combined remote sensing data on recent (2001–2014) disturbances with in-depth local information for 50 protected landscapes and their surroundings across the temperate biome. Disturbance patterns are highly variable, and shaped by variation in disturbance agents and traits of prevailing tree species. However, high disturbance activity is consistently linked to warmer and drier than average conditions across the globe. Disturbances in protected areas are smaller and more complex in shape compared to their surroundings affected by human land use. This signal disappears in areas with high recent natural disturbance activity, underlining the potential of climate-mediated disturbance to transform forest landscapes.Climate change may impact forest disturbances, though local variability is high. Here, Sommerfeld et al. show that disturbance patterns across the temperate biome vary with agents and tree traits, yet large disturbances are consistently linked to warmer and drier than average conditions.