The results of the Forest Resources Assessment 2000 carried out by FAO are synthetically presented and discussed. The world forest coverage is estimated equal to 38.6 million km 2 . The comparison of the estimates from the period 1990-2000 with those from the period 1980-1990 points out a certain general deceleration of the rate of net deforestation, that currently involves around 9 million hectares every year. However, the annua1 loss of tropical forests is still very large, while temperate and borea1 forests are in expansion. Overall, FRA2000 produced a relevant effort to compensate the existing technical, institutional and financial constraints and shortcomings for monitoring the world forest resources. The need to increase the quality and the frequency of forest surveys, both at national and international levels, s t a stands as a major issue to cope with.

http://ojs.aisf.it/index.php/ifm/article/download/774/738

Valutazione delle risorse forestali a livello globale

The huge amount of data currently produced by modern Earth Observation (EO) missions has allowed for the design of advanced machine learning techniques able to support complex Land Use/Land Cover (LULC) mapping tasks. The Copernicus programme developed by the European Space Agency provides, with missions such as Sentinel-1 (S1) and Sentinel-2 (S2), radar and optical (multi-spectral) imagery, respectively, at 10 m spatial resolution with revisit time around 5 days. Such high temporal resolution allows to collect Satellite Image Time Series (SITS) that support a plethora of Earth surface monitoring tasks. How to effectively combine the complementary information provided by such sensors remains an open problem in the remote sensing field. In this work, we propose a deep learning architecture to combine information coming from S1 and S2 time series, namely TWINNS (TWIn Neural Networks for Sentinel data), able to discover spatial and temporal dependencies in both types of SITS. The proposed architecture is devised to boost the land cover classification task by leveraging two levels of complementarity, i.e., the interplay between radar and optical SITS as well as the synergy between spatial and temporal dependencies. Experiments carried out on two study sites characterized by different land cover characteristics (i.e., the Koumbia site in Burkina Faso and Reunion Island, a overseas department of France in the Indian Ocean), demonstrate the significance of our proposal.

Combining Sentinel-1 and Sentinel-2 Satellite Image Time Series for land cover mapping via a multi-source deep learning architecture

Changes in vegetation activity are driven by multiple natural and anthropogenic factors, which can be reflected by Normalized Difference Vegetation Index (NDVI) derived from satellites. In this paper, NDVI trends from 1982 to 2012 are first estimated by the Theil–Sen median slope method to explore their spatial and temporal patterns. Then, the impact of climate variables and human activity on the observed NDVI trends is analyzed. Our results show that on average, NDVI increased by 0.46 × 10−3 per year from 1982 to 2012 globally with decadal variations. For most regions of the world, a greening (increasing)–browning (decreasing)–greening (G-B-G) trend is observed over the periods 1982–2004, 1995–2004, and 2005–2012, respectively. A positive partial correlation of NDVI and temperature is observed in the first period but it decreases and occasionally becomes negative in the following periods, especially in the Humid Temperate and Dry Domain Regions. This suggests a weakened effect of temperature on vegetation growth. Precipitation, on the other hand, is found to have a positive impact on the NDVI trend. This effect becomes stronger in the third period of 1995–2004, especially in the Dry Domain Region. Anthropogenic effects and human activities, derived here from the Human Footprint Dataset and the associated Human Influence Index (HII), have varied impacts on the magnitude (absolute value) of the NDVI trends across continents. Significant positive effects are found in Asia, Africa, and Europe, suggesting that intensive human activity could accelerate the change in NDVI and vegetation. A more accurate attribution of vegetation change to specific climatic and anthropogenic factors is instrumental to understand vegetation dynamics and requires further research.

Spatial and Temporal Patterns of Global NDVI Trends: Correlations with Climate and Human Factors

There is still large uncertainty over the status of global forest cover owing to the paucity of comprehensive and holistic studies related to long term forest cover change. The aim of the present work is to prepare a nation-wide multi-date forest cover database which describes and quantifies historical and recent changes in natural forests of India. This analysis facilitated the determination of the state of Indian forest cover changes over last eight decades. Here, we have mapped the total area under forest cover, evaluated the spatial tracking of changes in natural forests, estimated the rate of deforestation and afforestation, analysed the biogeographic zone wise and state wise forest cover change, existing land use in deforested area, influence of environmental factors such as terrain on deforestation and implication of different definitions of forest used by agencies reporting deforestation in India. The results indicated that forests covered an area of 869,012 km2 in 1930 which has decreased to 625,565 km2 in 2013, a net loss of 243,447 km2 (28 %) in eight decades. The highest annual average forest loss was found to be 4795 km2 during 1930–1975, 1476 km2 during 1975–1985, 767 km2 during 1985–1995, 356 km2 during 1995–2005 and 209 km2 during 2005–2013. Between 1930 and 1975, forest experienced large scale deforestation at gross annual rate of 0.77 % which has declined to 0.29 % and 0.14 % for the 1975–1985 and 1985–1995 periods respectively. Quantification of annual rate of gross deforestation for the recent period indicates 0.07 % during 1995–2005 and 0.05 % during 2005–2013. The lower rates of deforestation during recent period support effectiveness of conservation measures taken at national level. It was found that deforestation rate has decreased in many biogeographic zones by 2005, except for Andaman & Nicobar Islands and North East. The major deforestation has mostly occurred due to conversion of forests to agriculture. The construction of reservoirs contributed to 4.1 % of forest loss. The tropical forests have experienced large scale deforestation followed by subtropical forests. The findings of the study will be useful to prioritize conservation and protection of forest cover at the regional level. It shall also provide a base for future research on the impacts of deforestation on carbon flux and biodiversity.

Quantification and monitoring of deforestation in India over eight decades (1930–2013)

https://research.vu.nl/ws/files/121837637/Monitoring_biodiversity_in_the_Anthropocene_using_remote_sensing_in_species_distribution_models.pdf

Monitoring biodiversity in the Anthropocene using remote sensing in species distribution models

Understanding farmland abandonment in the Polish Carpathians

Broad scale forest cover reconstruction from historical topographic maps

Grassland use intensity is a topic of growing interest worldwide, as grasslands are integral in supporting biodiversity, food production, and regulating of the global carbon cycle. Data available for characterizing grasslands management are largely descriptive and collected from laborious field campaigns or questionnaires. The recent launch of the Sentinel-2 earth monitoring constellation provides new possibilities for high temporal and spatial resolution remote sensing data covering large areas. This study aims to evaluate the potential of a time series of Sentinel-2 data for mapping of mowing frequency in the region of Canton Aargau, Switzerland. We tested two cloud masking processes and three spatial mapping units (pixels, parcel polygons and shrunken parcel polygons), and investigated how missing data influence the ability to accurately detect and map grassland management activity. We found that more than 40% of the study area was mown before 15 June, while the remaining part was either mown later, or was not mown at all. The highest accuracy for detection of mowing events was achieved using additional clouds masking and size reduction of parcels, which allowed correct detection of 77% of mowing events. Additionally, we found that using only standard cloud masking leads to significant overestimation of mowing events, and that the detection based on sparse time series does not fully correspond to key events in the grass growth season.

/pdf/regional-scale-mapping-of-grassland-mowing-frequency-with-17sr0nerk4.pdf

Regional Scale Mapping of Grassland Mowing Frequency with Sentinel-2 Time Series

Global digital elevation models (DEMs) are an invaluable source of information in large area studies. Of particular interest are shuttle radar topography mission (SRTM) data that are freely available for the scientific community worldwide. Prior to any application, global datasets should be evaluated using reference data of higher accuracy. Therefore, the main objective of this study was to assess the accuracy of the SRTM C-band (version 4) DEM and SRTM X-band DEM of mountainous areas located in Poland and to examine the quality of data in relation to topographic parameters, radar beam geometry, initial voids in data and the presence of forest cover. A DEM from the Central National Geodetic and Cartographic Inventory, Poland, served as a reference. The study consisted of three steps: (i) the computation of vertical errors of the SRTM C- and X-band DEMs, (ii) the examination of any systematic bias in the data, and (iii) the analysis of the relationships between the elevation errors and terrain slope, aspect, local incidence angle, occurrence of voids and land cover. We found that the SRTM C- and X-band DEMs have mean errors equal to 4.31 ± 14.09 and 9.03 ± 37.40 m and root mean square errors equal to 14.74 and 38.47 m, respectively. Only 82 % of the C-band DEM and 74 % of the X-band DEM vertical errors had absolute values below 16 m. We found that the most important factors determining the occurrence of high errors were the distribution of initial voids and high slope angles for the C-band DEM, and local incidence angle, slope, aspect and radar beam geometry for the X-band DEM. In both cases, the presence of forest cover increased the mean error by approximately 10 m.

/pdf/assessment-of-the-accuracy-of-srtm-c-and-x-band-high-25mqrbj0zg.pdf

Assessment of the Accuracy of SRTM C- and X-Band High Mountain Elevation Data: a Case Study of the Polish Tatra Mountains

Secondary forest succession on abandoned agricultural land has played a significant role in land cover changes in Europe over the past several decades. However, it is difficult to quantify over large areas. In this paper, we present a conceptual framework for mapping forest succession patterns using vegetation structure information derived from LiDAR data supported by national topographic vector data. This work was performed in the Szczawnica commune in the Polish Carpathians. Using object-based image analysis segments of no vegetation, and sparse/dense low/medium/high vegetation were distinguished and subsequently compared to the national topographic dataset to delineate agricultural land that is covered by vegetation, which indicates secondary succession on abandoned fields. The results showed that 18.7% of the arable land and 40.4% of grasslands, that is 31.0% of the agricultural land in the Szczawnica commune, may currently be experiencing secondary forest succession. The overall accuracy of the approach was assessed using georeferenced terrestrial photographs and was found to be 95.0%. The results of this study indicate that the proposed methodology can potentially be applied in large-scale mapping of secondary forest succession patterns on abandoned land in mountain areas.

/pdf/mapping-secondary-forest-succession-on-abandoned-4meochip7h.pdf

Natalia Kolecka

Papers

Understanding farmland abandonment in the Polish Carpathians

Broad scale forest cover reconstruction from historical topographic maps

Regional Scale Mapping of Grassland Mowing Frequency with Sentinel-2 Time Series

Assessment of the Accuracy of SRTM C- and X-Band High Mountain Elevation Data: a Case Study of the Polish Tatra Mountains

Mapping Secondary Forest Succession on Abandoned Agricultural Land with LiDAR Point Clouds and Terrestrial Photography