The primary dataset used in this analysis was 5.405 GHz (wavelength, λ=5.5 cm; C-band) SAR imagery in ScanSAR wide mode at HH polarization from RADARSAT-2 acquired over the CAA (Fig. 1) in April from 2009–2018 (Table 1). RADARSAT-2 ScanSAR wide mode imagery has a spatial resolution of 100 m with an incidence angle range of 20.0 to 49.3∘. We limited our analysis to only RADARSAT-2 images at HH polarization because Scharien et al. (2017) found HV produced noisy results in addition to there not being sufficient imagery at HV polarization in the early period of the RADARSAT-2 record to cover CAA in April.

In situ observations of fp on landfast FYI were obtained in 2 consecutive years from sites in the CAA using a terrestrial light detection and ranging (lidar) system (Landy et al., 2014) (Fig. 1, green star). In 2011, the site was located in Allen Bay on FYI with relatively rough surface topography, whereas in 2012, the site was located in Resolute Passage on FYI with relatively smooth topography. At each site, time series of fp observations were collected within the same 100×100 m area of the ice over a 2- to 3-week period following melt onset, covering three of the four stages of melt pond evolution detailed in Eicken et al. (2002). The lidar system produces dense measurements over snow or sea ice, with specular reflection over melt ponds allowing melt pond fractions to be retrieved with an accuracy better than 5 % (Landy et al., 2014). These observations allow us to evaluate how well RADARSAT-2 resolves fpk of seasonally evolving sea ice coverage.

Aerial photographs of estimated fp directly over the lidar site and the adjacent sea ice area away from land and open water were also obtained on 22 June 2012. The aerial photographs have a 0.22 m pixel resolution, covering 750 m by 750 m. In total, 123 aerial photographs of fp were used, and a complete description of the dataset is provided in Scharien et al. (2014).

Finally, we made use of 8 d composite satellite observations of fp obtained from the MODIS Arctic melt pond cover fraction dataset that has a spatial resolution of 12.5 km for the period of 2009–2011 (Rösel et al., 2012) and weekly sea ice area and stage of development observations obtained from the Canadian Ice Service Digital Archive (CISDA) regional ice charts for the period of 2009–2018 (Tivy et al., 2011).

RADARSAT-2 fpk was determined using a modified approach to that described by Scharien et al. (2017). Their approach determines the second stage of the seasonal melt pond evolution cycle when fp is at its peak (Eicken et al., 2002; Polashenski et al., 2012) using Sentinel-1 Extra Wide (EW) swath imagery obtained during April within the CAA. April corresponds to late winter sea ice conditions in the CAA, when sea ice growth has reached its maximum and spring warming has yet to begin. Their approach was developed by relating the winter period HH gamma nought (γ∘) backscatter in decibels (dB) from Sentinel-1 to fpk observations in 1.7 m spatial resolution GeoEye-1 imagery, from spatially coincident image segments that represented homogeneous FYI and MYI regions. The result was that γ∘ can be converted to fpk using the following equation: 1fpk=-0.221-0.041γ∘. In Eq. (1), γ∘ was found to explain 73 % of the variability in fpk (Scharien et al., 2017).

In this study, all the available HH polarization RADARSAT-2 imagery over the CAA in April from 2009–2018 (Table 1) was first calibrated to γ∘ which minimizes the influence of incidence angle more so than with sigma nought (σ∘) (Small, 2011). RADARSAT-2 images were then speckle filtered using a 5×5 Lee filter and spatially registered to a common map projection. Finally, γ∘ was converted to fpk by applying Eq. (1) to each RADARSAT-2 image. For each year, the corresponding RADARSAT-2 fpk images in April were mosaicked together to cover the entire spatial domain of the CAA. Constructing a mosaic over a large region such as the CAA presents certain challenges with SAR imagery, particularly incidence angle variability. Even with the use of γ∘, Scharien et al. (2017) found that because of varying incidence angles associated with different ScanSAR images that fpk striping can still occur within the CAA in the mosaicked image. Our approach here was to average out incidence angle variability by taking advantage of a large amount of overlapping RADARSAT-2 imagery within the CAA (i.e. 90 to 159 images; Table 1) together with the fact that the majority of the sea ice in the CAA is landfast (immobile) during April, which results in a temporally stable fpk for all April images. To produce a RADARSAT-2 fpk mosaic within the CAA for each year, we calculated the mean fpk for each overlapping pixel using all of each year's RADARSAT-2 April images that effectively helped to reduce fpk striping across the CAA.

The root-mean-square error (RMSE) of fpk based on Eq. (1) is 0.085 (Scharien et al., 2017). While calculating the mean fpk of the overlapping image pixels helps reduce striping across the CAA, it also adds additional uncertainty and its effectiveness depends on the number of overlaps. In order to quantify the additional uncertainty (RMSER2), we used the mean and maximum standard deviation of RADARSAT-2 fpk of all pixels within the CAA calculated from 2009–2018 (fstd) together with a range of pixel overlaps (n) in the following equation: 2RMSER2=[(fstd/n0.5)2+0.0852]0.5. Since RADARSAT-2 imagery is acquired operationally, overlapping images vary interannually but pixel overlaps across the CAA were typically between 6–12. Figure 2 illustrates the RMSER2 values for a range of pixel overlaps using the 2009–2018 mean fstd value of 0.08 and the 2009–2018 maximum fstd value of 0.2. For the maximum fstd with pixel overlaps between 6–12 the RMSER2 ranges from 0.10–0.12.

The spatial distribution of mosaicked RADARSAT-2 fpk and pre-melt season (i.e. April) and sea ice stage of developed conditions in the CAA for the 2009–2018 time period are shown in Figs. 3 and 4, respectively. Lower fpk values are located primarily in the northern regions of the CAA (Queen Elizabeth Islands), Viscount Melville Sound, and the M'Clintock Channel where the majority of the CAA's MYI is typically found. The shallow bays and narrow channels located throughout the CAA exhibit high fpk values and these regions are typically associated with smooth FYI whereas rougher ice regions (i.e. Gulf of Boothia) are associated with lower fpk values. We should expect a lower fpk over MYI regions compared to FYI regions (Grenfell and Perovich, 2004; Perovich and Polashenski, 2012), and indeed the overall spatial distribution of RADARSAT-2 fpk is in excellent agreement with the spatial distribution of sea ice stage of development prior to the melt season for all years.

Figure 5a shows the time series of RADARSAT-2 fpk variability together with mean April MYI area in the CAA from 2009–2018. Over the 10-year record, the mean RADARSAT-2 fpk was 0.47 and ranged from a low of 0.43 in 2009 to a high of 0.52 in 2013. The temporal variability in RADARSAT-2 fpk is reflected in the variability of April MYI area within the CAA over the 10-year record with a statistically significant detrended correlation (R) of R=-0.89. The RADARSAT-2 fpk linkage with April MYI area is particularly evident from 2011 and 2012 which were very light sea ice years within the CAA whereby a considerable amount of the CAA's MYI area was lost during the summer melt season (Howell et al., 2013), and this resulted in 2012 and 2013 (i.e. the years following extreme melt) being the two highest RADARSAT-2 fpk years from 2009–2018 (Fig. 3d–e). MYI area within in the CAA then increased following these light ice years and RADARSAT fpk began to respond accordingly. In fact, there has always been a period of MYI recovery following light ice years with either MYI grown in situ and/or advected from Arctic Ocean into the CAA and gradually migrating to the CAA's southern regions (Howell et al., 2013). Figure 5b illustrates the standard deviation of RADARSAT-2 fpk from 2009–2018 and spatially reflects the process of MYI flowing southward through the CAA as RADARSAT-2 fpk was more variable in the MYI regions of the CAA compared to regions where FYI dominates the regional icescape.

What is interesting in Fig. 5a is that the mean RADARSAT-2 fpk in 2009 was lower than all years from 2014–2018 (with the exception of 2016) despite the CAA containing less MYI area. In addition, 2017 and 2018 exhibited a larger spatial coverage of MYI compared to 2009 (Fig. 4a, i–j). We suggest that higher RADARSAT-2 fpk in recent years is a result of Arctic Ocean MYI entering the CAA being younger and thinner than in 2009 (Howell and Brady, 2019) with smoother surface topography, thereby having a higher summer melt pond coverage (Landy et al., 2015). This seems to be particularly evident in the Viscount Melville Sound and M'Clintock Channels regions when comparing 2009 (Fig. 3a) with 2017 (Fig. 3i) and 2018 (Fig. 3j). Indeed, several studies have reported considerable decreases in the age and thickness of Arctic Ocean MYI north of the CAA in recent years (e.g. Kwok, 2018; Petty et al., 2020; Tschudi et al., 2020)

Frequency distributions of RADARSAT-2 fpk and Sentinel-1 fpk from Scharien et al. (2017) in the CAA for 2016 and 2017 are shown in Fig. 6. Sentinel-1 appears to estimate more regions of lower fpk compared to RADARSAT-2 which are typically associated with MYI. Whereas RADARSAT-2 estimates more regions of higher fpk which are typically associated with smooth FYI. We consider these subtle differences to be primarily the result of taking the mean of all available April RADARSAT-2 imagery (Table 1) over all incidence angles in the CAA compared to only using images from Sentinel-1 within the CAA constrained to a certain incident angle range. As shown in Fig. 2, the uncertainty in RADARSAT-2 fpk varies depending on the number of pixel overlaps (images). Overall, the fpk distributions are in good agreement between both sensors.

The in situ evolution of fp over FYI within the CAA acquired by Landy et al. (2014) allows us to place the RADARSAT-2 fpk estimates within the melt pond stages of development classification system. Unfortunately, no MODIS fp observations are located in close proximity to the in situ observations. The evolution of melt ponds on the surface of the sea ice has been classified into four distinct and consecutive stages. A brief description is provided here, and the reader is referred to Eicken et al. (2002) and Polashenski et al. (2012) for a more comprehensive description. In stage I, meltwater from snowmelt fills topographic depressions on the surface of the sea ice until the ponds reach their maximum areal extent. In stage II, melt pond coverage decreases due to horizontal water transport into macroscopic flaws and drainage through the ice. In stage III, the melt ponds typically drain through to the ocean and further changes in melt pond coverage depend on changes in surface topography and freeboard. Finally, in stage IV, melt ponds that survived the melt season refreeze and snow begins to accumulate on their surface.

Figure 7a compares the time series of the entire 100 m lidar melt pond fraction coincident with the fpk determined from RADARSAT-2 at the coinciding pixels. For 2011, RADARSAT-2 fpk corresponds to the end of stage I and beginning of stage II, thus providing a very good representation of the seasonal peak of the fp, when the melt pond control on heat uptake and ice decay, through the ice–albedo feedback, is greatest. For 2012, RADARSAT-2 fpk also corresponds to the end of stage I and beginning of stage II but is ∼ 0.2 lower than in situ fp values. This is likely due to the short duration but very high maximum fp of 0.78 in 2012 as Scharien et al. (2017) found that Eq. (1) sometimes underestimates very high fp due to the low γ∘ signal associated with very smooth FYI.

To give spatial context beyond the single-point comparison at the lidar site, Fig. 7b shows the distribution of RADARSAT-2 fpk and the fp determined from aerial photo observations on 22 June 2012 near Resolute. The aerial photographs were acquired within 1 week of fpk coverage being observed at the lidar site. The comparison was done by averaging all RADARSAT-2 pixels within each aerial photo (123 photos) which represents ∼ 861 samples. The mean aerial photograph fp was 0.54 and RADARSAT-2 fpk was 0.53 with an the RMSE of 0.10 and bias of 0. The distributions are in reasonably good agreement but RADARSAT-2 values are slightly narrower than the distribution of fp from the aerial photographs. It is likely the RADARSAT-2 distribution is narrow on the left tail because our method captures peak pond coverage, and some of the regions photographed were before or after their seasonal peak. We attribute the narrow right tail to the documented underestimation of Eq. (1) from Scharien et al. (2017). However, it is notable that both RADARSAT-2 and the aerial photograph datasets capture the same bimodal fp distribution, with the first mode around 0.4–0.5 characterizing rougher sea ice areas and the second mode around 0.7 capturing smooth flooded sea ice.

The seasonal time series of the 8 d composite MODIS fp, the maximum seasonal MODIS fp and the predicted RADARSAT-2 fpk for 2009–2011 is shown in Fig. 8. MODIS fp observations within the CAA indicate initial pond formation occurred in May for all years with fpk reached in mid-July for 2009 and in early June for 2010 and 2011. Compared to the RADARSAT-2 fpk values, the peak MODIS fp is ∼ 0.20 smaller. RADARSAT-2 fpk is higher on average than MODIS because the MODIS 8 d product does not represent fpk. The MODIS fp observations are determined weekly using 8 d composite image products that would include some melt pond formation and drainage processes prior to, and after, the seasonal peak. Moreover, MODIS fp values are essentially aggregated from 500 m clear-sky pixels within a 12.5 km × 12.5 km grid cell (Rösel et al., 2012) and the 500 m spatial resolution may limit detection of smaller pond fractions. Furthermore, not all of the 500 m pixels within the 12.5 km × 12.5 km grid cell are likely to be at the same melt pond stage evolution. Finally, MODIS fp observations give the time series of fp; therefore even the highest seasonal estimated MODIS fp is reduced because while some regions of the CAA are at their seasonal peak others are behind or ahead. To that end, we also calculated the maximum fp from MODIS regardless of timing during the melt season, for each pixel, also shown in Fig. 8. These values more closely compare with the RADARSAT-2 fpk but are still ∼ 0.05 smaller on average. Even the maximum fp from MODIS is from an 8 d running mean of daily pond fraction estimates, so will underestimate the fpk if the duration of peak ponding is < 8 d. However, the top whisker of the box plot of the maximum fp from MODIS indicates that MODIS does capture some regions at peak during the 8 d time series. Although we are using MODIS fp product to compare against our RADARSAT-2 fpk estimates, Rösel et al. (2012) found that the MODIS fp product also has errors up to ∼ 0.1. Overall, MODIS fp estimates are more representative of the seasonal mean fp rather than fpk within the CAA.

In order to investigate whether RADARSAT-2 fpk values can be used to provide predictive information for summer sea ice area within the CAA, we separated the CAA into numerous predefined subregions and then determined the detrended correlations between RADARSAT-2 fpk and weekly sea ice area from the CISDA regional ice charts in each region over the period of 2009–2018. We tested each week from the start of June to the end of September. The strongest correlation, together with the corresponding week of occurrence are shown in Fig. 9a and b, respectively. All the strongest correlations are negative, indicating – as expected – that years with higher predicted fpk values are associated with lower sea ice area at a later point in the summer. The higher fpk the lower the area-averaged albedo of the ice surface leading to accelerated melt and lower sea ice concentrations (e.g. Perovich and Polashenski, 2012). There is considerable spatial variability in the strongest correlation across the CAA with relatively low correlations in the majority of the northern CAA and very low correlations in the eastern regions of the CAA. The regions of Kellet-Crozier (R=-0.92), Viscount Melville Sound (R=-0.73), M'Clintock Channel (R=-0.77), and Norwegian Bay (R=-0.78) all exhibit statistically significant correlations above the 95 % confidence level. In terms of timing for the statistically significant regions, RADARSAT-2 fpk correlated the strongest to weekly sea ice area in August for all regions except Norwegian Bay (Fig. 9b). Compared to previous studies, the primary difference between using fp values to predict summer sea ice conditions seems to be the timing of when the correlation is the strongest. Using simulated fp values, Schröder et al. (2014) found the strongest correlation to September sea ice occurred for the May fp. Liu et al. (2015) used observed MODIS fp values and reported the strongest correlation to September sea ice in late July. Our findings suggest that methods such as these may be able to predict August sea ice area from fpk simulations or observations with higher confidence than September ice area, at least in the CAA.

Why is the relationship stronger in some regions of the CAA and weaker in others? RADARSAT-2 fpk values are determined from imagery acquired in April when ice conditions in the CAA are landfast (immobile) and do not evolve in concert with sea ice dynamics operating within the CAA. As a result, RADARSAT-2 fpk values will not be spatially representative of the region's ice conditions when region-specific dynamic breakup processes dominate over thermodynamics (i.e. in situ melt). In other words, the origin of the ice in these regions during the summer melt season will not always be the same as in April (i.e. pre-melt) when the initial RADARSAT-2 fpk value was determined. The time series of weekly detrended RADARSAT-2 fpk and weekly sea ice area for selected regions within the CAA is shown in Fig. 10 and provides evidence for this regional dichotomy. In the Viscount Melville Sound and M'Clintock regions the correlations gradually get stronger, reaching a peak in August. These regions are known to be immobile and stagnant (e.g. Melling, 2002) with the majority of breakup taking place in September, which is when the relationship begins to degrade. The Kellet-Crozier is another stagnant region, which supports that in the absence of considerable ice dynamics the relationship between RADARSAT-2 fpk and sea ice area is strong throughout the melt season. The time series in Penny Strait illustrates how the correlation gradually increases but when the region's dynamic breakup begins in July, ice is advected southward which degrades the correlation. This was also the case for many other regions in the northern CAA (not shown) as the flushing of sea ice southward from the northern CAA is a regular occurrence during the melt season (Melling, 2002; Howell et al., 2006). The low correlations in the southeastern regions of the CAA are also likely a function of ice dynamics as these regions of the CAA are known to be considerably influenced by currents and wind (Prinsenberg and Hamilton, 2005), and sea ice speed in Lancaster Sound and Barrow Strait can reach 10 km d-1 (Agnew et al., 2008).

The strong and statistically significant correlation in the Viscount Melville Sound region is encouraging as it is a key shipping region in the northern route of the Northwest Passage. To that end, we used linear regression to predict mean August sea ice area within Viscount Melville Sound with the detrended RADARSAT-2 fpk values as a predictor. Figure 11 illustrates the results compared to observations (detrended) from the CISDA ice charts for 2009–2018. There is reasonable agreement between the predicted and observed sea ice area in the region with an RMSE of 18×103 km2 and an R2=0.44. The largest discrepancies occurred for 2013 and 2014 with the RADARSAT-2 fpk model prediction resulting in too little sea ice area. Overall, within the Viscount Melville Sound region of CAA there is a period for which a significant statistical relationship exists between RADARSAT-2 fpk and the summer ice area before sea ice dynamics degrades the relationship.