Differencing of digital surface models derived from structure from motion (SfM) processing of airborne imagery has been used to produce snow depth (SD) maps with between ∼2 and ∼15 cm horizontal resolution and accuracies of ±10 cm over relatively flat surfaces with little or no vegetation and over alpine regions. This study builds on these findings by testing two hypotheses across a broader range of conditions: (i) that the vertical accuracy of SfM processing of imagery acquired by commercial low-cost unmanned aerial vehicle (UAV) systems can be adequately modelled using conventional photogrammetric theory and (ii) that SD change can be more accurately estimated by differencing snow-covered elevation surfaces rather than differencing a snow-covered and snow-free surface. A total of 71 UAV missions were flown over five sites, ranging from short grass to a regenerating forest, with ephemeral snowpacks. Point cloud geolocation performance agreed with photogrammetric theory that predicts uncertainty is proportional to UAV altitude and linearly related to horizontal uncertainty. The root-mean-square difference (RMSD) over the observation period, in comparison to the average of in situ measurements along ∼50 m transects, ranged from 1.58 to 10.56 cm for weekly SD and from 2.54 to 8.68 cm for weekly SD change. RMSD was not related to microtopography as quantified by the snow-free surface roughness. SD change uncertainty was unrelated to vegetation cover but was dominated by outliers corresponding to rapid in situ melt or onset; the median absolute difference of SD change ranged from 0.65 to 2.71 cm. These results indicate that the accuracy of UAV-based estimates of weekly snow depth change was, excepting conditions with deep fresh snow, substantially better than for snow depth and was comparable to in situ methods.

Sentinel-2 is a constellation of two satellites launched by the European Space Agency (ESA), respectively on 23 June 2015 and 7 March 2017, to map geophysical parameters over land surfaces. ESA provides Level 2 bottom-of-atmosphere reflectance (BOA) products (ESA-L2A) for Europe, with plans for operational global coverage, as well as the Sen2Cor (S2C) offline processor. In this study, aerosol optical thickness (AOT), precipitable water vapour (WVP) and surface reflectance from ESA-L2A products are compared with S2C output when using identical input Level 1 radiance products. Additionally, AOT and WVP are validated against reference measurement. As ESA and S2C share the same input and atmospheric correction algorithm, it was hypothesized that they should show identical validation performance and that differences between products should be negligible in comparison to the uncertainty of retrieved geophysical parameters due to radiometric uncertainty alone. Validation and intercomparison was performed for five clear-sky growing season dates for each of three ESA-L2A tiles selected to span a range of vegetation and topography as well as to be close to the AERONET measurement site. Validation of S2C (ESA) products using AERONET site measurements indicated an overall root mean square error (RMSE) of 0.06 (0.07) and a bias of 0.05 (0.09) for AOT and 0.20 cm (0.22 cm) and the bias was −0.02 cm (−0.10 cm) for WVP. Intercomparison of S2C-L2A and ESA-L2A showed an overall agreement higher than 99% for scene classification (SCL) maps and negligible differences for WVP (RMSE under 0.09 and R² above 0.99). Larger disagreement was observed for aerosol optical thickness (AOT) (RMSE up to 0.04 and R² as low as 0.14). For BOA reflectance, disagreement between products depends on vegetation cover density, topography slope and spectral band. The largest differences were observed for red-edge and infrared bands in mountainous vegetated areas (RMSE up to 4.9% reflectance and R² as low as 0.53). These differences are of similar magnitude to the radiometric calibration requirements for the Sentinel 2 imager. The differences had minimal impact of commonly used vegetation indices (NDVI, NDWI, EVI), but application of the Sentinel Level 2 biophysical processor generally resulted in proportional differences in most derived vegetation parameters. It is recommended that the consistency of ESA and S2C products should be improved by the developers of the ESA and S2C processors.

Time series of vegetation biophysical variables (leaf area index (LAI), fraction canopy cover (FCOVER), fraction of absorbed photosynthetically active radiation (FAPAR), canopy chlorophyll content (CCC), and canopy water content (CWC)) were estimated from interpolated Sentinel-2 (S2-LIKE) surface reflectance images, for an agricultural region located in central Canada, using the Simplified Level 2 Product Prototype Processor (SL2P). S2-LIKE surface reflectance data were generated by blending clear-sky Sentinel-2 Multispectral Imager (S2-MSI) images with daily BRDF-adjusted Moderate Resolution Imaging Spectrometer images using the Prediction Smooth Reflectance Fusion Model (PSFRM), and validated using thirteen independent S2-MSI images (RMSE $\le$ 6%). The uncertainty of S2-LIKE surface reflectance data increases with the time delay between the prediction date and the closest S2-MSI image used for training PSFRM. Vegetation biophysical variables from S2-LIKE products are validated qualitatively and quantitatively by comparison to the corresponding vegetation biophysical variables from S2-MSI products (RMSE = 0.55 for LAI, ~10% for FCOVER and FAPAR, and 0.13 g/m² for CCC and 0.16 kg/m² for CWC). Uncertainties of vegetation biophysical variables derived from S2-LIKE products are almost linearly related to the uncertainty of the input reflectance data. When compared to the in situ measurements collected during the Soil Moisture Active Passive Validation Experiment 2016 field campaign, uncertainties of LAI (0.83) and FCOVER (13.73%) estimates from S2-LIKE products were slightly larger than uncertainties of LAI (0.57) and FCOVER (11.80%) estimates from S2-MSI products. However, equal uncertainties (0.32 kg/m²) were obtained for CWC estimates using SL2P with either S2-LIKE or S2-MSI input data.

Snow depth (SD) can vary by more than an order of magnitude over length scales of metres due to topography, vegetation and microclimate. Differencing of digital surface models derived from Structure from Motion (SfM) processing of airborne imagery has been used to produce SD maps with between ∼2 cm to ∼15 cm horizontal resolution and accuracies on the order of ±10 cm over both relatively flat surfaces with little or no vegetation and over alpine regions. Studies indicate that accuracy is lower in the presence of vegetation above or below the snowpack and in rough topography; suggesting that some biases may be temporally persistent. Moreover, flight and image parameters vary across studies but they are typically not related a priori to an expected uncertainty in SD. This study tests two hypotheses: i) that SD change can be more accurately estimated when differencing snow covered elevation surfaces rather than the absolute snow depth based on differencing a snow covered and snow free surface and ii) the vertical accuracy of SfM processing of imagery acquired by commercial light weight unmanned aerial vehicle (UAV) systems can be adequately modelled using conventional photogrammetric theory. Moreover, these hypotheses are tested over areas with ephemeral snow pack conditions and across a range of micro-topography and vegetation cover. Weekly SD maps with

The Sentinel 2 (S2) constellation mission was designed to facilitate the systematic mapping canopy biophysical variables at medium resolution on a global basis and in a free and open manner.  The mission concept requires the development of downstream services to map variables such as the fraction of absorbed photosynthetically active radiation (fAPAR), fraction of canopy cover (fCOVER) and leaf area index (LAI) using Level 2A surface reflectance inputs from the S2 ground segment.  Currently, free and open products generation can be performed using the Simplified Level 2 Prototype Processor (SL2P) applied on a product granule basis.  Considering that the processor is a prototype this study addresses three questions: 1) Can the SL2P algorithm, or subsequent versions, be engineered to facilitate systematic product generation over large extents in a free and open manner? 2) What is the uncertainty of SL2P products over North America during the growing season? 3) Can the uncertainty be reduced by changing the calibration database used within SL2P?  To facilitate validation and product generation, SL2P was ported to a Google Earth Engine application (the Landscape Evolution and Forecasting Toolbox).  This now allows mapping of up to one million square kilometers in near real time using either the original SL2P algorithm or updated versions.  SL2P uncertainty was quantified over North America using direct comparison to 20 in-situ sites within the National Environmental Observing Network in the continental United States of America and within a Canada wide field campaign over forests and shrublands conducted by Canada Centre for Remote Sensing. SL2P outputs were also compared to MODIS and Copernicus Global Land Service products over the Belmanip II regional sites and 30 additional forested regions in North America.  Results from NEON validation indicate SL2P is generally within uncertainty requirements except for forests; where it underestimates fAPAR, fCOVER and LAI.  Results for other sites will also be presented.  To address the forest bias, SL2P was recalibrated using simulations from the FLIGHT 3D radiative transfer model representative of North American forests.  The uncertainty of the recalibrated SL2P algorithm will be compared to baseline SL2P estimates to determine if increased model complexity is warranted.