o Equally high sensitivity of coherence and HV- backscatter intensity to forest parameters. While these results indicate that L-band data are most suitable for forest cover mapping and forest parameter retrieval, the development of related algorithms for large-scale applications has to take into account the availability of SAR datasets as well. The ERS mission lasts since 1991 and has produced an enormous archive of SAR images. Multi-temporal approaches were found to increase the usefulness of C-band backscatter datasets for forestmapping  and retrieval of forest parameters . During the ERS-1/2 tandem mission in 1995-1999 several image pairs with one-day temporal separation have been acquired, boosting the use of coherence in forest-related studies. Simple classification approaches and relatively straightforward retrieval techniques provided results with high accuracy . Nonetheless, the coverage of the northernhemisphere is only partial. Although the Envisat mission lasting since 2002 can be seen as follow-up of the ERS mission, it misses the short-term interferometric component (repeat-pass of 35 days). Systematic and large-scale acquisitions take place in the low-resolution ScanSAR mode only (150 m and 500 m). The very high density of observations within a short time period has revived the use of the multi- temporal approaches developed for ERS images . With data from the JERS-1 mission (1992-1998), it was primarily proven that (i) large-scale forestmapping with radar is feasible , (ii) retrieval of forest parameters is possible with a limited set of measurements  and (iii) coherence is exploitable despite the long temporal baseline . The legacy of the JERS-1 mission has served to set up the observation strategy of the ALOS PALSAR mission. This can be considered as the currently most suitable system for forest observations with radar, thanks also to the polarimetric capability. The PALSAR mission however misses in part the multi- temporal aspect, which we believe is the key for deriving more detailed information on forests with respect to one or a small number of observations. 3. FOREST COVER MAPPING
which do not necessarily - or only temporally - exhibit the backscatter properties of forest floor (e.g. Dobson et al., 1995b; Kasischke et al., 1997; 2007; Brisco & Brown, 1998; Proisy et al., 2000; Quegan et al., 2000; Moran et al., 2000; Henderson & Lewis, 2008). Concerning soil moisture, the soil moisture levels at the forest floor and open unforested ground can differ because of the effects of evapotranspiration and rainfall interception over forest (Geiger et al., 2003). This could explain why the VCF-based model training tended to overestimate σ 0 gr in case of unfrozen and rainy weather conditions. Accordingly, the observation that the VCF-based estimates for σ 0 gr were consistently better for the winter images may be explained with the frozen snow covered ground that diminishes a lot of the potential differences between forest floor and areas without tree cover. As in case of intensity, one may expect that it could be critical to infer from the coherence observed over non-forest land on the coherence at the forest floor. However, the comparison of the inventory data- and VCF-based estimates for γ gr (when restricting the VCF-based training to the area of the particular test site) revealed less pronounced differences than in case of σ 0 gr . A reason for this could have been that coherence, in contrast to intensity, is not sensitive to soil moisture differences that may arise between forest and open land because of, for instance, evapotranspiration or rainfall interception. Only heterogeneous variations of soil moisture between the tandem acquisitions lower the coherence (Luo et al., 2001). Not much knowledge has been gathered yet concerning the question when soil moisture fluctuations are heterogeneous or homogeneous at the sub-pixel scale though. This may be related to the root concentration, soil type/porosity or ground water depth. In addition, the impact of understorey vegetation on coherence has not been analyzed yet. It is thus difficult to appraise under which conditions it may be critical to infer from the coherence that was measured in areas without tree cover on the forest floor coherence. As in case of intensity, it can be assumed that potential differences should be minimized in case of frozen ground surfaces. Nevertheless, the ERS-1/2 tandem pairs that were acquired over Bolshe-Murtinsky showed that, in particular under unfrozen and wet conditions, the ground coherence level can vary considerably within the area of an ERS frame but the variations were obviously a consequence of soil type specific effects and did not reflect systematic differences in coherence between forest floor and unforested terrain.
This study presents results for observing forest changes in Sweden using multi-temporal L- band satellite data and is a part of the JAXA’s ALOS Kyoto and Carbon Initiative. An extensive dataset of images acquired by the Advanced Land Observing Satellite Phased Array type L-band Synthetic Aperture Radar (ALOS PALSAR) is investigated for clear- cut detection in boreal forests in northern Sweden (Lat. 64°14’ N, Long. 19°50’ E). Strong forest/non-forest contrast and temporal consistency were found for the Fine Beam Dual HV-polarized backscatter during unfrozen conditions. Thus, a simple thresholding algorithm that exploits the temporal consistency of pair-wise HV-backscatter measurements has been developed for detection of clear-felled areas. When applied to an image pair acquired during favorable weather conditions, the detection algorithm identified 76% of the clear-cut pixels within a reference layer, with zero erroneously detected pixels. With further refinement, the developed methodology can be an option to present operational alternatives for clear-cut detection.
Large-footprint lidar systems ( Blair et al., 1999 ) have been developed to provide high-resolution, geo-located measurements of
vegetation vertical structure and ground elevations beneath dense canopies. Over the past decade, several airborne and space-borne large-footprint lidar systems have been used to make measurements of vegetation. The lidar waveform signature from large-footprint lidar instrument, such as the Scanning Lidar Imager of Canopies by Echo Recovery (SLICER) ( Harding et al., 1995, 1998 ) and the Laser Vegetation Imaging Sensor (LVIS) ( Blair et al., 1999 ) has been successfully used to estimate the tree height and forest above-ground biomass ( Drake et al., 2002, 2003; Dubayah & Drake, 2000; Hofton et al., 2002; Lefsky et al., 1999a,b; Sun et al., 2008 ). The relationship between forest carbon storage and the vertical structure from lidar waveform is relatively unexplored. Further studies on the data properties, (e.g. the effects of multiple scattering and ground slope on lidar signatures) are needed to verify and improve the retrieval algorithms. One major limitation of current spaceborne lidar systems (i.e., ICESat GLAS) is the lack of imaging capabilities and the fact that they provide sparse sampling information on the forest structure.
This study allowed end users to obtain geohazard mapping tools that define the spatial extent, temporal variation, amplitude and rate of ground movement. 4 GEOTECHNICAL APPLICATIONS, PERMAFROST Satellite based InSAR monitoring is applicable to several specific permafrost conditions. In the northern regions unique geocryological processes such as frost heave, settlement caused by permafrost thawing, thermokarst, solifluction, and slope processes commonly occur. These processes cause vertical and/or horizontal ground displacements which in turn pose significant challenges for design and safe operation of structure and infrastructure facilities such as buildings, dams, roads and pipelines. These vertical and/or horizontal ground displacements can be monitored and verified through satellite based InSAR monitoring. Such processes can start during the construction phase and continue into the service phase. Geosystems in polar conditions can be very sensitive to even very small changes in geological conditions, because of its metastable balance. Aberration of natural geosystem balance can trigger these negative geocryological processes.
1. I NTRODUCTION
Due to the capability of acquiring high-resolution radar images independently of weather conditions and day-night cycle, spaceborne and airborne Synthetic Aperture Radar (SAR) systems have been effectively used in numerous Earth observation applications more than thirty years. Military applications such as targeting, surveillance and exploration; environmental monitoring of polar ice, glacier, ocean currents, oil spill, vegetation, soil, floods; and coherent applications such as change detection and elevation modeling have been the main SAR applications. Remote sensing of forest is economically significant for today’s world and also becoming more significant due to the acceleration in climate and ecosystem changes. In forestry, most SAR studies focus on retrieving forest parameters related to the 3D structure such as tree height, biomass, vertical and horizontal heterogeneity as well as the intrinsic properties such as tree type, moisture content and leaf area index from the observed quantities such as backscattering coefficient and scattering phase center height.
A low point density LiDAR dataset (LD) with small footprint (35 cm) was acquired in 1996 during an airborne geophysical survey that covered 4/5 of French Guiana. Unlike the waveform data acquired by GLAS, this LiDAR dataset provides only the first return pulse. Thus, the data acquired corresponds to the elevation of the first object encountered by the laser beam. Moreover, the database contains laser elevations every 7 m, along track, on flight lines spaced 500 m apart and oriented 30 ˝ N, intersected by transverse flight lines spaced 5 km apart and oriented at 120 ˝ N. The estimation of canopy heights using the LD dataset showed that canopy heights reached a mean canopy height of 30 m. Only 1% of canopy heights were higher than 50 m in all of French Guiana. In order to estimate canopy heights from the LD dataset, a three-step procedure was implemented [ 24 ]. First, LiDAR points acquired over canopy tops were identified (called top-of-canopy points). Next, LiDAR points acquired between two consecutive top-of-canopies were identified (called pseudo-ground points), and then the lowest pseudo-ground point in a distance of 1 km was identified as the ground point. Finally, for each ground point and corresponding top-of-canopy point the canopy height was estimated by calculating the distance from the ground point to the segment joining the two top-of-canopy points. Over French Guiana, canopy height estimates from the LD dataset had a density of 1.19 points/km 2 . Low canopy heights (maximum of 20 m) can be observed in the northern parts of French Guiana on the coastal marsh areas.
While the existence of a strange attractor for climate has proven difficult and is limited by a number of factors, the aims of this study are to search for a finite attractor for a specific climate subsystem, represented by extra- tropical NorthernHemisphere stratospheric variability, and to study the dependence of the dimension of the attractor on the choice of the analyzed variable, using four different reanalyses. Results are also tested for two model runs with different forcings in order to test the dependence of the dimension of the attractor to the seasonality of the forcing. The analysis of the time series yields a convergence to finite dimensions for all the different variables and datasets. Convergence of the dynamical entropies can however only be shown for a subset of variables and datasets. The different dimen- sions of the attractor follow the hierarchy of dimensions D l # D 2 # D 1 # D 0 # D KY . Interestingly, however, results show a large difference in values between the correlation (D 2 ) and the Lyapunov (D KY ) dimensions, suggesting the presence of a multifractal structure for the attractor associated with the variability of the system (Parisi and Frisch 1985; Halsey et al. 1986).
27 6. Discussion
6.1 Evaluation of method and accuracy of the result
All the field observations received for this study had to be used as input into the change detection analysis, which is why no empirical map accuracy assessment could be performed. The lack of field measurement was because it was not possible to design and conduct field measurements in the timeframe of this study or to be in situ of the active feeding period of the geometrids. However, I argue that the created maps have high accuracy for several reasons; (1) Threshold values ranged from -1.397 to 1.386 for non-infested forest and <-2.0 for highly likely infested forest. A precautionary principle was used to determine the high likely infested threshold, so there would be no overlap with the non-infested forest class. By having an interval between these two opposite classes, it minimized the chance for non-infested forest to be classified incorrectly. (2) The curve fitted to raw seasonal NDVI data, smoothed by the TIMESAT program, kept its characteristics and effectively reduced noise, as illustrated in figure 6 with the very close fit to the raw data. (3) Bylund and Landström who visited the area during the infestation of 2012 confirmed the infestation classification as correct in regards to the places they had visited, and according to their knowledge and judgments of the distribution (email communication with Bylund and Landström). (4) Outbreaks can be expected to follow a similar pattern of distribution if they occur consecutively. The 2012 outbreak followed a similar pattern of distribution as that found for 2004 by Babst et al. (2009). During both outbreaks the north side of the lake was not as heavily infested as the south.
(bare) habitats may be a prerequisite to formation of obligate serpentinophytes (Armbruster 2014; Cacho and Strauss 2014) and it is likely that in the northern lati- tudes this process is currently in its early stages, where subspecies, ecotypes, or races diﬀering in edaphic toler- ance or in their ability to hyperaccumulate Ni have yet to evolve as full-ﬂedged species (Brummitt et al. 1987; Nyberg Berglund et al. 2004; Brysting 2008; Teptina and Paukov 2015). Territories in the Holarctic which have never been frozen bear high numbers of these species—at least 215 ultramaﬁc endemic taxa are known in Cali- fornia, and the Mediterranean region is likewise rich in endemics (Anacker 2011). The only obligate ser- pentinophyte Alyssum litvinovii Knjaz. (Fig. 5b) is cur- rently known from the Southern Urals but in the territory outside the circumboreal region (Knjasev 2011). Arid territories of Holarctic which have never been aﬀected by glaciation processes may therefore represent a unique opportunity to discover new taxa which may qualify as obligate to ultramaﬁc soils. These territories may include Mugodzhary mountains in Kazakhstan, Caucasus and Altai and are all worthy of intense ﬁeld exploration.
Abstract. We analyse the spatio-temporal patterns of tem- perature variability over NorthernHemisphere land areas, on centennial time-scales, for the last 12 centuries using an unprecedentedly large network of temperature-sensitive proxy records. Geographically widespread positive temper- ature anomalies are observed from the 9th to 11th centuries, similar in extent and magnitude to the 20th century mean. A dominance of widespread negative anomalies is observed from the 16th to 18th centuries. Though we find the ampli- tude and spatial extent of the 20th century warming is within the range of natural variability over the last 12 centuries, we also find that the rate of warming from the 19th to the 20th century is unprecedented in the context of the last 1200 yr. The positive NorthernHemisphere temperature change from the 19th to the 20th century is clearly the largest between any two consecutive centuries in the past 12 centuries. These results remain robust even after removing a significant num- ber of proxies in various tests of robustness showing that the choice of proxies has no particular influence on the overall conclusions of this study.
Improved understanding of NorthernHemisphere paleoto- pography of intermediate-sized ice sheets can, if combined with the relevant array of paleoclimatic proxy data, yield a valuable calibration data set for tuning and comparison of climate models at global ice volumes smaller than the LGM case normally used (Kageyama et al., 1999; Otto-Bliesner et al., 2006). The present situation, in which research focus has been on two climatic end members, largely precludes re- search on non-linear responses to ice sheet build-up. Over the last decades evidence has accumulated that ice sheet extent and volume varied more drastically over shorter time scales than previously thought, and a new paradigm of highly dy- namic ice sheets has emerged (Boulton and Clark, 1990). Of particular importance is the evidence for ice-free core ar- eas in Fennoscandia (Wohlfarth, 2010; Helmens and Engels, 2010), the consequence of which is that the Fennoscandian Ice Sheet had to be fully rebuilt from zero ice volume dur- ing each of the cold phases (MIS 5d, 4 and 2). From this
Radar altimeters are nadir-looking instruments that determine the distance or range from the sensor to the scattering surface by transmitting and receiving radar pulses. The range and spatial resolution of a radar altimeter are governed by the same principles as described in Section 3.1. Radar altimeters are, however, mostly operated in pulse-limited mode, that is the footprint on the ground is determined by the pulse length rather than the opening angle of the antenna. For distributed targets such as ocean and ice surfaces, the measured distance corresponds to a mean surface of all the scatterers within the footprint area. The resulting measurement is therefore susceptible to speckle noise and several pulses are averaged to obtain an accurate height measurement. This averaged signal is called a waveform and reflects received power depending on range from the antenna. A schematic of the observation geometry, the pulse limited footprint and the received waveform is shown on the left of Figure 4.1. The waveform shows an initial rise when the leading edge of the pulse is reflected by the surface and reaches its maximum when the trailing edge of the pulse is first reflected from the surface. The illuminated area at that time instance is a circle (depicted in blue on the left of Figure 4.1) and equivalent to the pulse limited footprint. Thereafter the illuminated area forms annuli of equal area. Theoretically the waveform would remain constant but decreases primarily because of antenna pattern weighting [76–78]. The pulse limited footprint of such conventional spaceborneradar altimeters is in the order of 1.5–2.5 km in diameter but can be significantly larger for rough seas [79, 80]. The beam-limited footprint is determined by the opening angle of the antenna and is on the order of 15 km.
The methodology in this study is based on the merging of LiDAR canopy height estimates (airborne (LD) and spaceborne) with ancillary data. The spaceborne LiDAR dataset corresponds to data acquired by the GLAS sensor with a density of 0.51 points/km². The canopy heights produced from this dataset showed a precision of 3.6 m . The airborne LiDAR data (LD) corresponds to data acquired in 1996 over almost 80% of the total study area with a point density of 1.19 points/km² and a precision on the canopy height estimates of about 1.5 m in comparison to a high density LiDAR dataset (HD) . Finally, the ancillary maps are a geological map (GEO), a forest landscape type map (LT), maps derived from the shuttle topography mission (SRTM) data (slope, terrain surface roughness (Rug), and drain map (ln_drain)), mean rainfall map from the past 10 years derived from data provided by the NASA tropical rainfall measuring mission (Rain), and three maps derived from Enhanced vegetation index (EVI). These maps represent the minimum EVI value from the last 10 years (EVI_MIN), the average EVI value (EVI_AVG), and the maximum EVI value (EVI_MAX).
Evidence from satellite observation. In situ phenology obser- vations cover only a small fraction of world’s vegetation types, geographic ranges and climate gradients. To evaluate the generality of the in situ-observed asymmetric temperature effects on leaf onset in Europe and parts of the US, we further analysed the effects of daytime and nighttime temperature changes on satellite-derived VGD in the terrestrial northernhemisphere (4 30°N) over the past 30 years (1982–2011; see Methods). VGD at 0.5 0.5° resolution was estimated from time series of the NDVI3g data set (1982–2011) developed by the Global Inventory Modeling and Mapping Studies (GIMMS) group (see Methods). Note the reported VGD here is the average value from four different VGD algorithms: Spline Midpoint, Hants Maximum, Polyﬁt Maximum and Timesat SG (Savitzky–Golay; see Methods). Similar to the in situ-observation results, satellite- derived L max ranged between 0 and 3 months across 76% of the
European storm Kyrill, which impacted Western, Central, and Eastern Europe in 2007 brought over 1 billion dollars in damage after intensifying over the northern Atlantic (Fink et al 2009). With considerable impact comes the need to understand the formation and evolution of both types of cyclones, along with their role in a changing climate, in order to minimize loss of life and property and further the overall understanding involved in forecasting these events. In a previous study by Bengtsson et al (2006) on extratropical cyclones in a warming climate, an ensemble of three 30-yr integrations of the Max Planck Institute (MPI) coupled atmosphere– ocean model (OM; ECHAM OM) was used to compare storm tracks for the period 2070–99 with the control period 1960–89. They found that there was a general poleward shift in cyclone tracks during this period with little to no intensification. This is in agreement with other previous studies that utilized similar Lagrangian techniques studies such as Yin (2005) and Fischer-Bruns et al. (2005). Within this paper, intensification will be a reference to either the depth of the pressure field as estimated by the mean sea level pressure (MSLP) or the magnitude of the relative vorticity field at the 850 millibar level. This is in contrast to fields such as wind speed and precipitation, which are arguably equally important, particularly from the perspective of the public.
atmosphere could be disrupted by a major nuclear war in the NorthernHemisphere. During the northern spring and summer, air rises over the hot humid Tropics, splits into two streams and descends over the subtropical and middle latitudes of both hemispheres to create secondary areas of air circulation at higher latitudes (diagram at left). If a large, dense cloud of smoke and dust were to be introduced into the lower atmosphere in the northern temperate latitudes during these seasons, the heating at the southern edge of the cloud might be intense enough to reverse the normal mid-latitude descent of air and create an unusual air flow pattern in which upper-level winds blow briskly across the equator from north to south (diagram at right). Adapted from Scientific American, Vol. 251 (August 1984).
Here we give a brief description of the data used and then an overview of the steps involved in the methodology before describing each individually in detail. We produced storm tracks for all NorthernHemisphere cyclones between 2008 and 2012 from the ECMWF NWP operational analysis. This is a high-resolution dataset produced through the process of data assimilation. The TCs were then identified and selected by matching against the best track record (Hodges and Emerton, 2014). The point in the CPS occupied by the storm tracks at each 6 hourly (i.e. synoptic time) timestep was then determined. As such the path through phase space of each storm was determined. All the CPS points for all storms in the dataset were then clustered according to an unsupervised K-means clustering algorithm. The properties of these clusters, such as number of members and their location in the phase space, were then investigated to determine the cluster, or clusters, that represented ET. These were then used to classify the life- cycle progression of individual storms, thus objectively determining ET.
[ 17 ] This study has investigated the contribution of extra-
tropical cyclones to the total precipitation climatology of the NorthernHemisphere. Decomposing how important different phenomena are to the total precipitation climatology provides valuable insight into understanding how precipitation may respond to climate change. Using an objective feature tracking algorithm, extratropical cyclones have been identified in the NorthernHemisphere. The precipitation associated with these cyclones has been assessed in order to provide an estimate of how important the contribution of these cyclones is to the total precipitation climatology. The regions which are identified as having the greatest absolute storm associated precipitation are in the storm tracks of the Atlantic and the Pacific. It is also shown that in many regions, including parts of Europe and much of North America, over 70% of total precipitation is associated with the passage of an extratropical cyclone.