MODIS Terra Collection 6 fractional snow cover validation in mountainous terrain during spring snowmelt using Landsat TM and ETM+

Daily swath MODIS Terra Collection 6 fractional snow cover (MOD10_L2) estimates were validated with two‐day Landsat TM/ETM + snow‐covered area estimates across central Idaho and southwestern Montana, USA. Snow cover maps during spring snowmelt for 2000, 2001, 2002, 2003, 2005, 2007, and 2009 were compared between MODIS Terra and Landsat TM/ETM + using least‐squared regression. Strong spatial and temporal map agreement was found between MODIS Terra fractional snow cover and Landsat TM/ETM + snow‐covered area, although map disagreement was observed for two validation dates. High‐altitude cirrus cloud contamination during low snow conditions as well as late season transient snowfall resulted in map disagreement. MODIS Terra's spatial resolution limits retrieval of thin‐patchy snow cover, especially during partially cloudy conditions. Landsat's image acquisition frequency can introduce difficulty when discriminating between transient and resident mountain snow cover. Furthermore, transient snowfall later in the snowmelt season, which is a stochastic accumulation event that does not usually persist beyond the daily timescale, will skew decadal snow‐covered area variability if bi‐monthly climate data record development is the objective. As a quality control step, ground‐based daily snow telemetry snow‐water‐equivalent measurements can be used to verify transient snowfall events. Users of daily MODIS Terra fractional snow products should be aware that local solar illumination and sensor viewing geometry might influence fractional snow cover estimation in mountainous terrain. Cross‐sensor interoperability has been confirmed between MODIS Terra and Landsat TM/ETM + when mapping snow from the visible/infrared spectrum. This relationship is strong and supports operational multi‐sensor snow cover mapping, specifically climate data record development to expand cryosphere, climate, and hydrological science applications. Copyright © 2013 John Wiley & Sons, Ltd.     

Uncertainty in seasonal snow reconstruction: Relative impacts of model forcing and image availability

There are many areas of uncertainty when solving the inverse problems of snow water equivalent (SWE) reconstruction. These include (i) the ability to infer the Final Date of the Seasonal Snow (FDSS) cover, particularly from remote sensing; (ii) errors in model forcing data (such as air temperature or radiation fluxes); and (iii) weaknesses in the snow model used for the reconstruction, associated with both the fidelity of the equations used to simulate snow processes (structural uncertainty) and the parameter values selected for use in the model equations. We investigate the trade-offs among these sources of uncertainty using 10,000 station-years worth of data from the western US SNOTEL network. Model structural and parameter uncertainty are eliminated by using a perfect model scenario i.e. comparing results to modelled control runs. The model was calibrated for each station-year to ensure that the model simulations reflect reality. Results indicate that for a temperature index model, a ±5 days error in FDSS gives a median −25%/+32% error in maximum SWE. A 1 °C air temperature bias produces a SWE error larger than a 5 days error in the FDSS for 50% of the 10,000 cases. Similarly, a 5 days error in FDSS could be accounted for by a net radiation error of 13 W m⁻² or less during the melt period, in 50% of cases. Mean absolute errors of 1 °C or more are typically reported in the literature for air temperature interpolations at high elevations. Observed solar radiation during the melt season can differ by 30 W m⁻² over relatively short distances, while estimates from reanalysis (NARR, ERA-Interim, MERRA, CFSRR) and GOES satellites typically span more than 40 W m⁻². Using data from both MODIS sensors (Terra & Aqua) at all snow covered points in the western US, a consecutive 5 days gap in imagery at time of FDSS is likely to occur only 5–10% of the time. This work shows that errors in model forcing data are at least as important, if not more, than image availability when reconstructing SWE.     

Accuracy assessment of the MODIS snow products

A suite of Moderate‐Resolution Imaging Spectroradiometer (MODIS) snow products at various spatial and temporal resolutions from the Terra satellite has been available since February 2000. Standard products include daily and 8‐day composite 500 m resolution swath and tile products (which include fractional snow cover (FSC) and snow albedo), and 0·05° resolution products on a climate‐modelling grid (CMG) (which also include FSC). These snow products (from Collection 4 (C4) reprocessing) are mature and most have been validated to varying degrees and are available to order through the National Snow and Ice Data Center. The overall absolute accuracy of the well‐studied 500 m resolution swath (MOD10_L2) and daily tile (MOD10A1) products is ～93%, but varies by land‐cover type and snow condition. The most frequent errors are due to snow/cloud discrimination problems, however, improvements in the MODIS cloud mask, an input product, have occurred in ‘Collection 5’ reprocessing. Detection of very thin snow (<1 cm thick) can also be problematic. Validation of MOD10_L2 and MOD10A1 applies to all higher‐level products because all the higher‐level products are all created from these products. The composited products may have larger errors due, in part, to errors propagated from daily products. Recently, new products have been developed. A fractional snow cover algorithm for the 500 m resolution products was developed, and is part of the C5 daily swath and tile products; a monthly CMG snow product at 0·05° resolution and a daily 0·25° resolution CMG snow product are also now available. Similar, but not identical products are also produced from the MODIS on the Aqua satellite, launched in May 2002, but the accuracy of those products has not yet been assessed in detail. Published in 2007 by John Wiley & Sons, Ltd.     

A comparison of MODIS and NOHRSC snow‐cover products for simulating streamflow using the Snowmelt Runoff Model

Remote sensing is an important source of snow‐cover extent for input into the Snowmelt Runoff Model (SRM) and other snowmelt models. Since February 2000, daily global snow‐cover maps have been produced from data collected by the Moderate Resolution Imaging Spectroradiometer (MODIS). The usefulness of this snow‐cover product for streamflow prediction is assessed by comparing SRM simulated streamflow using the MODIS snow‐cover product with streamflow simulated using snow maps from the National Operational Hydrologic Remote Sensing Center (NOHRSC). Simulations were conducted for two tributary watersheds of the Upper Rio Grande basin during the 2001 snowmelt season using representative SRM parameter values. Snow depletion curves developed from MODIS and NOHRSC snow maps were generally comparable in both watersheds: satisfactory streamflow simulations were obtained using both snow‐cover products in larger watershed (volume difference: MODIS, 2·6%; NOHRSC, 14·0%) and less satisfactory streamflow simulations in smaller watershed (volume difference: MODIS, −33·1%; NOHRSC, −18·6%). The snow water equivalent (SWE) on 1 April in the third zone of each basin was computed using the modified depletion curve produced by the SRM and was compared with in situ SWE measured at Snowpack Telemetry sites located in the third zone of each basin. The SRM‐calculated SWEs using both snow products agree with the measured SWEs in both watersheds. Based on these results, the MODIS snow‐cover product appears to be of sufficient quality for streamflow prediction using the SRM in the snowmelt‐dominated basins. Copyright © 2005 John Wiley & Sons, Ltd.     

Evaluation of snow products over the Tibetan Plateau

Four satellite‐based snow products are evaluated over the Tibetan Plateau for the 2007–2010 snow seasons. The Moderate Resolution Imaging Spectroradiometer (MODIS) Terra and Aqua snow cover daily L3 Global 500‐m grid products (MOD10A1 and MYD10A1), the National Oceanic and Atmospheric Administration Interactive Multisensor Snow and Ice Mapping System (IMS) daily Northern Hemisphere snow cover product and the Advanced Microwave Scanning Radiometer – Earth Observing System Daily Snow Water Equivalent were validated against Thematic Mapper (TM) snow cover maps of Landsat‐5 and meteorological station snow depth observations. The overall accuracy of MOD10A1, MYD10A1 and IMS is higher than 91% against stations observations and than 79% against Landsat TM images. In general, the daily MODIS snow cover products show better performance than the multisensor IMS product. However, the IMS snow cover product is suitable for larger scale (~4km) analysis and applications, with the advantage over MODIS to allow for mitigation for cloud cover. The accuracy of the three products decreases with decreasing snow depth. Overestimation errors are most common over forested regions; the IMS and Advanced Microwave Scanning Radiometer – Earth Observing System Snow Water Equivalent products also show poorer performance that the MODIS products over grassland. By identifying weaknesses in the satellite products, this study provides a focus for the improvement of snow products over the Tibetan plateau. The quantitative evaluation of the products proposed here can also be used to assess their relative weight in data assimilation, against other data sources, such as modelling and in situ measurement networks. Copyright © 2015 John Wiley & Sons, Ltd.     

Response of snow cover and runoff to climate change in high Alpine catchments of Eastern Switzerland

A model study on the impact of climate change on snow cover and runoff has been conducted for the Swiss Canton of Graubünden. The model Alpine3D has been forced with the data from 35 Automatic Weather Stations in order to investigate snow and runoff dynamics for the current climate. The data set has then been modified to reflect climate change as predicted for the 2021–2050 and 2070–2095 periods from an ensemble of regional climate models.The predicted changes in snow cover will be moderate for 2021–2050 and become drastic in the second half of the century. Towards the end of the century the snow cover changes will roughly be equivalent to an elevation shift of 800 m. Seasonal snow water equivalents will decrease by one to two thirds and snow seasons will be shortened by five to nine weeks in 2095.Small, higher elevation catchments will show more winter runoff, earlier spring melt peaks and reduced summer runoff. Where glacierized areas exist, the transitional increase in glacier melt will initially offset losses from snow melt. Larger catchments, which reach lower elevations will show much smaller changes since they are already dominated by summer precipitation.     

Simulating wind-affected snow accumulations at catchment to basin scales

In non-forested mountain regions, wind plays a dominant role in determining snow accumulation and melt patterns. A new, computationally efficient algorithm for distributing the complex and heterogeneous effects of wind on snow distributions was developed. The distribution algorithm uses terrain structure, vegetation, and wind data to adjust commonly available precipitation data to simulate wind-affected accumulations. This research describes model development and application in three research catchments in the Reynolds Creek Experimental Watershed in southwest Idaho, USA. All three catchments feature highly variable snow distributions driven by wind. The algorithm was used to derive model forcings for Isnobal, a mass and energy balance distributed snow model. Development and initial testing took place in the Reynolds Mountain East catchment (0.36 km²) where R² values for the wind-affected snow distributions ranged from 0.50 to 0.67 for four observation periods spanning two years. At the Upper Sheep Creek catchment (0.26 km²) R² values for the wind-affected model were 0.66 and 0.70. These R² values matched or exceeded previously published cross-validation results from regression-based statistical analyses of snow distributions in similar environments. In both catchments the wind-affected model accurately located large drift zones, snow-scoured slopes, and produced melt patterns consistent with observed streamflow. Models that did not account for wind effects produced relatively homogenous SWE distributions, R² values approaching 0.0, and melt patterns inconsistent with observed streamflow. The Dobson Creek (14.0 km²) application incorporated elevation effects into the distribution routine and was conducted over a two-dimensional grid of 6.67 × 10⁵ pixels. Comparisons with satellite-derived snow-covered-area again demonstrated that the model did an excellent job locating regions with wind-affected snow accumulations. This final application demonstrated that the computational efficiency and modest data requirements of this approach are ideally suited for large-scale operational applications.     

A comparison between two statistical and a physically-based model in snow water equivalent mapping

Snow water equivalent (SWE) estimates at the end of the winter season have been compared for the 2002–2006 period in a 200 km² mountainous area in Switzerland, using three different models. The first model, ALPINE3D, is a physically based process-oriented model, which solves the snowpack energy and mass balance equations. The other two models, SWE-SEM and HS-SWE, are statistical algorithms interpolating snow data on a grid. While SWE-SEM interpolates local estimates of SWE, HS-SWE converts interpolated snow depth maps into maps of SWE using a regionally-calibrated conversion model. We discuss similarities and differences among the models’ results, both in terms of total volume, and spatial distribution of SWE. The comparison shows a general good agreement of the results of the three models, with a mean difference in the total volumes between the two statistical models of ∼8%, and between the physical model and the statistical ones of ∼−3% to −10%.     

Automatic mapping of snow cover depletion curves using optical remote sensing data under conditions of frequent cloud cover and temporary snow

Snow cover depletion curves are required for several water management applications of snow hydrology and are often difficult to obtain automatically using optical remote sensing data owing to both frequent cloud cover and temporary snow cover. This study develops a methodology to produce accurate snow cover depletion curves automatically using high temporal resolution optical remote sensing data (e.g. Terra Moderate Resolution Imaging Spectroradiometer (MODIS), Aqua MODIS or National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR)) by snow cover change trajectory analysis. The method consists of four major steps. The first is to reclassify both cloud‐obscured land and snow into more distinct subclasses and to determine their snow cover status (seasonal snow cover or not) based on the snow cover change trajectories over the whole snowmelt season. The second step is to derive rules based on the analysis of snow cover change trajectories. These rules are subsequently used to determine for a given date, the snow cover status of a pixel based on snow cover maps from the beginning of the snowmelt season to that given date. The third step is to apply a decision‐tree‐like processing flow based on these rules to determine the snow cover status of a pixel for a given date and to create daily seasonal snow cover maps. The final step is to produce snow cover depletion curves using these maps. A case study using this method based on Terra MODIS snow cover map products (MOD10A1) was conducted in the lower and middle reaches of the Kaidu River Watershed (19 000 km²) in the Chinese Tien Shan, Xinjiang Uygur Autonomous Region, China. High resolution remote sensing data (charge coupled device (CCD) camera data with 19·5 m resolution of the China and Brazil Environmental and Resources Satellite (CBERS) data (19·5 m resolution), and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data with 15 m resolution of the Terra) were used to validate the results. The study shows that the seasonal snow cover classification was consistent with that determined using a high spatial resolution dataset, with an accuracy of 87–91%. The snow cover depletion curves clearly reflected the impact of the variation of temperature and the appearance of temporary snow cover on seasonal snow cover. The findings from this case study suggest that the approach is successful in generating accurate snow cover depletion curves automatically under conditions of frequent cloud cover and temporary snow cover using high temporal resolution optical remote sensing data. Copyright © 2007 John Wiley & Sons, Ltd.     

Estimation of snow density using full-polarimetric Synthetic Aperture Radar (SAR) data

In this paper, a new snow density estimation methodology is proposed for full-polarimetric Synthetic Aperture Radar (SAR) data. The generalized four component polarimetric decomposition with unitary transformation (G4U) based generalized volume parameter is utilized to invert snowpack dielectric constant using the Fresnel transmission coefficients. The snow density is then estimated using an empirical relationship. Six Radarsat-2 fine resolution full-polarimetric C-band datasets were acquired over Himachal Pradesh, India. The near-real time in-situ measurements were collected with the satellite pass to validate the proposed method. The mean absolute error (MAE) of the proposed method is 0.027 g cm⁻³ and the root mean square error (RMSE) is 0.032 g cm⁻³. The snow density variation within a season were also analyzed using multi-temporal Radarsat-2 data.     

Snow model sensitivity analysis to understand spatial and temporal snow dynamics in a high‐elevation catchment

In this paper, we addressed a sensitivity analysis of the snow module of the GEOtop2.0 model at point and catchment scale in a small high‐elevation catchment in the Eastern Italian Alps (catchment size: 61 km²). Simulated snow depth and snow water equivalent at the point scale were compared with measured data at four locations from 2009 to 2013. At the catchment scale, simulated snow‐covered area (SCA) was compared with binary snow cover maps derived from moderate‐resolution imaging spectroradiometer (MODIS) and Landsat satellite imagery. Sensitivity analyses were used to assess the effect of different model parameterizations on model performance at both scales and the effect of different thresholds of simulated snow depth on the agreement with MODIS data. Our results at point scale indicated that modifying only the “snow correction factor” resulted in substantial improvements of the snow model and effectively compensated inaccurate winter precipitation by enhancing snow accumulation. SCA inaccuracies at catchment scale during accumulation and melt period were affected little by different snow depth thresholds when using calibrated winter precipitation from point scale. However, inaccuracies were strongly controlled by topographic characteristics and model parameterizations driving snow albedo (“snow ageing coefficient” and “extinction of snow albedo”) during accumulation and melt period. Although highest accuracies (overall accuracy = 1 in 86% of the catchment area) were observed during winter, lower accuracies (overall accuracy < 0.7) occurred during the early accumulation and melt period (in 29% and 23%, respectively), mostly present in areas with grassland and forest, slopes of 20–40°, areas exposed NW or areas with a topographic roughness index of ?0.25 to 0 m. These findings may give recommendations for defining more effective model parameterization strategies and guide future work, in which simulated and MODIS SCA may be combined to generate improved products for SCA monitoring in Alpine catchments.     

Landsat TM image segmentation for delineating geological zone correlated vegetation stratification in the Kruger National Park,South Africa

Image classification approaches are widely used in mapping vegetation on remotely sensed images. Vegetation assemblages are equivalent to habitats. Whereas sub-pixel classification approaches potentially can produce more realistic, homogenous habitat maps, pixel-based hard classifier approaches often result in non-homogenous habitat zones. This salt-and-pepper habitat mapping is particularly a challenge on images of savannas, given the characteristic patchy texture of scattered trees and grass. Image segmentation techniques offer possibilities for homogenous habitat classification. This study aimed at establishing the extent to which established, field surveyed and geology-related vegetation types in South Africa’s Kruger National Park (KNP) can be reproduced using image segmentation. Rain season Landsat TM images were used, selected to coincide with the peak in vegetation productivity, which was deemed the time of year when discrimination between key habitats in KNP is most likely to be successful. The multiresolution segmentation mode in eCognition 5.0 was employed, object classification accomplished using the nearest neighbour (NN) classifier, using object texture and training area mean values in the NN feature space.Compared to delineations of the vegetation types of KNP on a digital map of the vegetation zones that was tested, image segmentation successfully mapped the zones (overall accuracy 85.3%, K^ = 82.7%) despite slight shifts in the location of vegetation zone boundaries. Maximum likelihood classification (MLC) of the same images was only 37% accurate (K^ = 24.2%). Whereas the vegetation zones resulting from MLC were non-homogenous, with considerable spectral confusion among the vegetation zones, image segmentation produced more homogenous vegetation zones, comparably more useful for conservation management, because realistic and meaningful habitat maps are important in biodiversity conservation as input data upon which to base management decisions. Image segmentation appears to be a useful approach in mapping savanna vegetation.     

Evaluation of the MODIS snow cover fraction product

Eleven years of daily 500 m gridded Terra Moderate Resolution Imaging Spectroradiometer (MODIS) (MOD10A1) snow cover fraction (SCF) data are evaluated in terms of snow presence detection in Colorado and Washington states. The SCF detection validation study is performed using in‐situ measurements and expressed in terms of snow and land detection and misclassification frequencies. A major aspect addressed in this study involves the shifting of pixel values in time due to sensor viewing angles and gridding artifacts of MODIS sensor products. To account for this error, 500 m gridded pixels are grouped and aggregated to different‐sized areas to incorporate neighboring pixel information. With pixel aggregation, both the probability of detection (POD) and the false alarm ratios increase for almost all cases. Of the false negative (FN) and false positive values (referred to as the total error when combined), FN estimates dominate most of the total error and are greatly reduced with aggregation. The greatest POD increases and total error reductions occur with going from a single 500 m pixel to 3×3‐pixel averaged areas. Since the MODIS SCF algorithm was developed under ideal conditions, SCF detection is also evaluated for varying conditions of vegetation, elevation, cloud cover and air temperature. Finally, using a direct insertion data assimilation approach, pixel averaged MODIS SCF observations are shown to improve modeled snowpack conditions over the single pixel observations due to the smoothing of more error‐prone observations and more accurately snow‐classified pixels. Copyright © 2012 John Wiley & Sons, Ltd.     

An evaluation of methods for determining during-storm precipitation phase and the rain/snow transition elevation at the surface in a mountain basin

Determining surface precipitation phase is required to properly correct precipitation gage data for wind effects, to determine the hydrologic response to a precipitation event, and for hydrologic modeling when rain will be treated differently from snow. In this paper we present a comparison of several methods for determining precipitation phase using 12 years of hourly precipitation, weather and snow data from a long-term measurement site at Reynolds Mountain East (RME), a headwater catchment within the Reynolds Creek Experimental Watershed (RCEW), in the Owyhee Mountains of Idaho, USA. Methods are based on thresholds of (1) air temperature (T_a) at 0 °C, (2) dual T_a threshold, −1 to 3 °C, (3) dewpoint temperature (T_d) at 0 °C, and (4) wet bulb temperature (T_w) at 0 °C. The comparison shows that at the RME Grove site, the dual threshold approach predicts too much snow, while T_a, T_d and T_w are generally similar predicting equivalent snow volumes over the 12 year-period indicating that during storms the cloud level is at or close to the surface at this location. To scale up the evaluation of these methods we evaluate them across a 380 m elevation range in RCEW during a large mixed-phase storm event. The event began as snow at all elevations and over the course of 4 h transitioned to rain at the lowest through highest elevations. Using 15-minute measurements of precipitation, changes in snow depth (z_s), T_a, T_d and T_w, at seven sites through this elevation range, we found precipitation phase linked to the during-storm surface humidity. By measuring humidity along an elevation gradient during the storm we are able to track changes in T_d to reliably estimate precipitation phase and effectively track the elevation of the rain/snow transition during the event.     

Simulated watershed responses to land cover changes using the Regional Hydro‐Ecological Simulation System

In this work, we used the Regional Hydro‐Ecological Simulation System (RHESSys) model to examine runoff sensitivity to land cover changes in a mountain environment. Two independent experiments were evaluated where we conducted simulations with multiple vegetation cover changes that include conversion to grass, no vegetation cover and deciduous/coniferous cover scenarios. The model experiments were performed at two hillslopes within the Weber River near Oakley, Utah watershed (USGS gauge # 10128500). Daily precipitation, air temperature and wind speed data as well as spatial data that include a digital elevation model with 30 m grid resolution, soil texture map and vegetation and land use maps were processed to drive RHESSys simulations. Observed runoff data at the watershed outlet were used for calibration and verification. Our runoff sensitivity results suggest that during winter, reduced leaf area index (LAI) decreases canopy interception resulting in increased snow accumulations and hence snow available for runoff during the early spring melt season. Increased LAI during the spring melt season tends to delay the snow melting process. This delay in snow melting process is due to reduced radiation beneath high LAI surfaces relative to low LAI surfaces. The model results suggest that annual runoff yield after removing deciduous vegetation is on average about 7% higher than with deciduous vegetation cover, while annual runoff yield after removing coniferous vegetation is on average as about 2% higher than that produced with coniferous vegetation cover. These simulations thus help quantify the sensitivity of water yield to vegetation change. Copyright © 2013 John Wiley & Sons, Ltd.     

Simulated effects of cropland expansion on seasonal temperatures over China

Human activities result in deforestation, expansion of cropland, grassland degradation, urbanization and other large-scale land use/cover change; among these, cropland expansion is one of the most important processes. To understand the effects of cropland expansion on seasonal temperatures over China, two 21-year simulations (spanning January 1, 1980–December 31, 2000), using the Regional Integrated Environmental Model System (RIEMS 2.0), were performed. The two simulations comprised current realistic land use/cover patterns and the previous vegetation cover without crop expansion, to investigate the impact of crop expansion on seasonal temperatures over China. The results showed that due to cropland expansion: (1) the most obvious changes occurred in the maximum temperatures, followed by the mean surface air temperatures, and the minimum temperatures were the least affected; (2) the summer mean maximum temperatures decreased in most parts of eastern China, and the temperatures changed significantly in most parts of northeast China, north China and central China (p < 0.05); (3) the surface air temperatures, maximum temperatures and minimum temperatures in summer decreased in the different regions by between −0.03 and −0.76 °C (the greatest temperature changes occurred in southwest China, and the smallest were in northeast China); (4) the net radiation flux and latent heat flux increased, while the sensible flux decreased, when semi-desert vegetation was replaced by dry land crops, in both summer and winter seasons, and the converse occurred when irrigated crops were replaced by dry land crops. In addition, the net radiation flux and sensible heat flux decreased, and the latent heat flux increased when short grass and tall grass were replaced dry land crops, as well as when dry land crops were replaced by irrigated crops.     

Evaluating the utility of the EUMETSAT HSAF snow recognition product over mountainous areas of eastern Turkey

Abstract

Monitoring snow parameters (e.g. snow-cover area, snow water equivalent) is challenging work. Because of its natural physical properties, snow strongly affects the evolution of weather on a daily basis and climate on a longer time scale. In this paper, the snow recognition product generated from the MSG-SEVIRI images within the framework of the Hydrological Satellite Facility (HSAF) Project of EUMETSAT is presented. Validation of the snow recognition product H10 was done for the snow season (from 1 January to 31 March) of the water year 2009. The MOD10A1 and MOD10C2 snow products were also used in the validation studies. Ground truth of the products was obtained by using 1890 snow depth observations from 20 meteorological stations, which are mainly located in mountainous areas and are distributed across the eastern part of Turkey. The possibility of 37% cloud cover reduction was obtained by merging 15-min observations from MSG-SEVIRI as opposed to using only one daily observation from MODIS. The coarse spatial resolution of the H10 product gave higher commission errors compared to the MOD10A1 product. Snow depletion curves obtained from the HSAF snow recognition product were compared with those derived from the MODIS 8-day snow cover product. The preliminary results show that the HSAF snow recognition product, taking advantage of using high temporal frequency measurement with spectral information required for snow mapping, significantly improves the mapping of regional snow-cover extent over mountainous areas.

Citation Surer, S. and Akyurek, Z., 2012. Evaluating the utility of the EUMETSAT HSAF snow recognition product over mountainous areas of eastern Turkey. Hydrological Sciences Journal, 57 (8), 1–11.     

Evaluation of DEM generation based on Interferometric SAR using TanDEM-X data in Tokyo

This study is focused on the evaluation of a Digital Elevation Model (DEM) for Tokyo, Japan from data collected by the recently launched TerraSAR add-on for Digital Elevation Measurements (TanDEM-X), satellite of the German Aerospace Center (DLR). The aim of the TanDEM-X mission is to use Interferometric SAR techniques to generate a consistent high resolution global DEM dataset. In order to generate an accurate global DEM using TanDEM-X data, it is important to evaluate the accuracy at different sites around the world. Here, we report our efforts to generate a high-resolution DEM of the Tokyo metropolitan region using TanDEM-X data. We also compare the TanDEM-X DEM with other existing DEMs for the Tokyo region. Statistical techniques were used to calculate the elevation differences between the TanDEM-X DEM and the reference data. Two high-resolution LiDAR DEMs are used as independent reference data. The vertical accuracy of the TanDEM-X DEM evaluated using the Root Mean Square Error (RMSE) is considerably higher than the existing global digital elevation models. However, the local area DEM generated by Geospatial Information Authority of Japan (GSI DEM) showed the highest accuracy among all non-LiDAR DEM’s. The vertical accuracy in terms of RMSE estimated using the 2 m LiDAR as reference is 3.20 m for TanDEM-X, 2.44 m for the GSI, 7.00 m for SRTM DEM and 10.24 m for ASTER-GDEM. We also compared the accuracy of TanDEM-X with the other DEMs for different types of land cover classes. The results show that the absolute elevation error of TanDEM-X is higher for urban and vegetated areas, likewise to those observed for other global DEM’s. This is probably because the radar signals used by TanDEM-X tend to measure the first reflective surface that is encountered, which is often the top of the buildings or canopy. Hence, the TanDEM-X based DEM is more akin to a Digital Surface Model (DSM).     

Using machine learning to produce near surface soil moisture estimates from deeper in situ records at U.S. Climate Reference Network (USCRN) locations: Analysis and applications to AMSR-E satellite validation

Surface soil moisture is a critical parameter for understanding the energy flux at the land atmosphere boundary. Weather modeling, climate prediction, and remote sensing validation are some of the applications for surface soil moisture information. The most common in situ measurement for these purposes are sensors that are installed at depths of approximately 5 cm. There are however, sensor technologies and network designs that do not provide an estimate at this depth. If soil moisture estimates at deeper depths could be extrapolated to the near surface, in situ networks providing estimates at other depths would see their values enhanced. Soil moisture sensors from the U.S. Climate Reference Network (USCRN) were used to generate models of 5 cm soil moisture, with 10 cm soil moisture measurements and antecedent precipitation as inputs, via machine learning techniques. Validation was conducted with the available, in situ, 5 cm resources. It was shown that a 5 cm estimate, which was extrapolated from a 10 cm sensor and antecedent local precipitation, produced a root-mean-squared-error (RMSE) of 0.0215 m³/m³. Next, these machine-learning-generated 5 cm estimates were also compared to AMSR-E estimates at these locations. These results were then compared with the performance of the actual in situ readings against the AMSR-E data. The machine learning estimates at 5 cm produced an RMSE of approximately 0.03 m³/m³ when an optimized gain and offset were applied. This is necessary considering the performance of AMSR-E in locations characterized by high vegetation water contents, which are present across North Carolina. Lastly, the application of this extrapolation technique is applied to the ECONet in North Carolina, which provides a 10 cm depth measurement as its shallowest soil moisture estimate. A raw RMSE of 0.028 m³/m³ was achieved, and with a linear gain and offset applied at each ECONet site, an RMSE of 0.013 m³/m³ was possible.     

Land use and land cover classification using phenological variability from MODIS vegetation in the Upper Pangani River Basin,Eastern Africa

In arid and semi-arid areas, evaporation fluxes are the largest component of the hydrological cycle, with runoff coefficient rarely exceeding 10%. These fluxes are a function of land use and land management and as such an essential component for integrated water resources management. Spatially distributed land use and land cover (LULC) maps distinguishing not only natural land cover but also management practices such as irrigation are therefore essential for comprehensive water management analysis in a river basin. Through remote sensing, LULC can be classified using its unique phenological variability observed over time. For this purpose, sixteen LULC types have been classified in the Upper Pangani River Basin (the headwaters of the Pangani River Basin in Tanzania) using MODIS vegetation satellite data. Ninety-four images based on 8 day temporal and 250 m spatial resolutions were analyzed for the hydrological years 2009 and 2010. Unsupervised and supervised clustering techniques were utilized to identify various LULC types with aid of ground information on crop calendar and the land features of the river basin. Ground truthing data were obtained during two rainfall seasons to assess the classification accuracy. The results showed an overall classification accuracy of 85%, with the producer’s accuracy of 83% and user’s accuracy of 86% for confidence level of 98% in the analysis. The overall Kappa coefficient of 0.85 also showed good agreement between the LULC and the ground data. The land suitability classification based on FAO-SYS framework for the various LULC types were also consistent with the derived classification results. The existing local database on total smallholder irrigation development and sugarcane cultivation (large scale irrigation) showed a 74% and 95% variation respectively to the LULC classification and showed fairly good geographical distribution. The LULC information provides an essential boundary condition for establishing the water use and management of green and blue water resources in the water stress Pangani River Basin.