Ultraviolet Radiation and Bio-optics in Crater Lake, Oregon, 2005
Long-term Proxies for Kd,UV and Decadal Changes in UV Transparency
Several proxies for UV attenuation can be used to look at long term changes in UV transparency because of the long time period covered by the primary measurements. The signal most similar to Kd,UV is Kd,blue measured by photometers equipped with changeable filters on deck and underwater sensors. LI-COR LI-1800uw radiometer data averaged over depths 0–40 m and for the period 1996–2002 were used to calculate Kd plotted in Figure 14A. These were used to derive regression equations (see Figure 14A caption) relating Kd,blue (nominally 400–500 nm) to Kd,320 and Kd,380 (1996–2002). We then applied the equations to historic Kd,blue data (gathered by photometers equipped with blue filters in 1940, 1969 and 1980–1991) to estimate Kd,320 and Kd,380. Figure 14B shows the time series for Kd,320 and Kd,380 derived from direct UV measurements (black symbols), from Kd,blue measurements with blue-filter photometers (light gray squares and triangles), and from radiometer measurements of only Kd,380 values (dark gray triangles) where Kd,320 was calculated using the relationship in Figure 6. Based on this collection of direct and indirect calculations the lowest Kd, 320 (0.066 m-1 estimated from radiometer-measured Kd,380) occurred in 1966; Kd, 320 in 2001 was nearly identical (0.068 m-1). The range among the annual summer averages for Kd,320 during this period (0.07 to 0.12 m-1) represents approximately 35% variation above or below the mean.
The longest optical record in Crater Lake is from measurements of Secchi depth, beginning in 1896 (Larson et al., 1996a). The inverse of Secchi depth (1/ZSD) should be correlated with other apparent and inherent optical properties as described earlier. When Kd,blue is calculated from blue-filter photometer data and averaged from the surface to the Secchi depth measured on the same day, 13 years of summer averages (June–August) are well-correlated with 1/ZSD (Figure 15A, Kd,blue = 2.08/ZSD – 0.026; r2=0.54, N=13). Figure 15B compares Kd,blue calculated from 1/ZSD for the period 1937 through 2002 with Kd,blue measured by photometer (averaged to ZSD or 40 m) and by radiometer (averaged to 40 m). The agreement among Kd,blue summer averages (July–August) from Photometer and 1/ZSD data (regardless of whether data were collected on the same day) was within +20% and -28% for the 12 years where both two types of data were available. Comparing Kd,blue derived from radiometer and ZSD the agreement was within +7% and -35%. The largest difference between radiometer and 1/ZSD averages occurred in 2000 when the summer averages were 0.046 m-1 from all 1/ZSD data and 0.031 m-1 from the two radiometer dates (June and August). The agreement improved when only 1/ZSD data collected within 2 days of the radiometer profiles were compared (agreement of summer averages was then within 1.5%). Differences can also be expected in part because instrument profiles covered 0–40 m while Secchi depth summer averages ranged from 23–38 m during the period 1896–2002.
Using the relationships described above for estimating Kd,blue from Secchi depth measurements and for estimating Kd,320 and Kd,380 from Kd,blue, we calculated Kd,320 and Kd,380 for the period 1896–2002. The resulting record of UV attenuation (averaged from the surface to depths ranging from 23–44 m, depending on the Secchi depth) is shown in Figure 16A from 1937–2002 along with radiometer-derived and photometer-derived values. In years when LI-1800uw radiometer or photometer records coincide with Secchi depth data there is a general correspondence between estimates except for a few years (early 1980s and 2000). The large spike in 1995 is based only on ZSD measurements although cp660 data confirm the high attenuation near the surface (Figures 9A & 9B). The range for Kd,320 based on Secchi depths is 0.08–0.19 m-1 (maximum 0.15 excluding 1995); based on the blue photometer data it is 0.08–0.15 m-1 (no measurements in 1995); based on the LI-1800uw it is 0.07–0.10 m-1 (1996–2002). From the detailed measurements starting in the 1978 there is a decadal pattern of peaks and valleys with amplitude of roughly +/- 50%, scattered with infrequent spikes of higher attenuation (e.g. 1995). By averaging all sources of data for Kd,320 for each year (Figure 16B) the range and pattern from 1896–1969 was plotted. The values ranged from 0.07–0.14 m-1 from 1978–2002, and from 0.07–0.12 m-1 from 1896–1969.