1 Evaluation of ClimateBC Dave Spittlehouse, BC Min Forests, Lands and Natural Resource Operations, Victoria, BC & Tongli Wang, Centre for Forest Conservation Genetics, Department of Forest and Conservation Sciences, University of British Columbia, Vancouver BC 5 February 214
2 Summary ClimateBC is a software page that produces high spatial resolution estimates of temperature and precipitation and a number of derived climate variables for historic and future climates. ClimateBC is an update of ClimateBC V4 and the BC component of ClimateWNA. Changes include a new base layer of PRISM monthly temperature and precipitation at 8 m resolution for the 1971-2 period, downscaling to specific elevations using lapse rates based on surrounding grid cells, a solar radiation base layer, CMIP5 climate change projections for three normals periods, and monthly time series of selected climate projections. This report evaluates ClimateBC s ability to produce historic monthly climate data. Calculated values are compared to measured climate normals and derived variables (1951-8, 1961-9 and 1971- periods) for specific locations (Meteorological Service of Canada weather stations) and to estimates from the previous version of the software. A spatial comparison over BC of ClimateBC and V4 was made for selected variables. Differences in climate of the zones and variants of BC s Biogeoclimatic (BEC) vegetation classification system calculated by Climate BC and ClimateWNA were determined. The testing evaluates the accuracy of the PRISM base layer in combination with the downscaling methodology and the CRU anomaly data used to adjust to different time period than the base period. The new 1971- PRISM layers at 8 m grid provide a slight improvement over the old 1961-9 PRISM layers at 4 km grid in predicting monthly temperature and precipitation of the measured climate of a weather station within the grid cell. As expected, downscaling to the station elevation further improves this prediction. However, the improvements between V4 and are marginal for temperature (using R 2 as an indicator) and there is a slight degradation on precipitation. Bias appears to be reduced for temperature and precipitation. These results are reflected in the derived variables, such as degree days, heat:moisture index and evaporation, which show only small or no improvement. The strength of PRISM is its estimates of temperature and precipitation in areas far removed from points of measurement. Maps of BC for the difference between ClimateBC and ClimateBC V4 for selected variables were generated at 8 m resolution. In general, the new PRISM results in a warmer interior for BC at high elevations and cooler for the coast. This cooler coastal area was noted in testing of individual station data where it was found that frost free period was underestimated at four lighthouses. There are some areas of northern BC that are cooler. Part of this is due to the fact that a splined data set rather than PRISM was used for northeastern BC in earlier versions of the software. Histogram plots show that most of the estimates of mean annual temperature fall within ± 1 C of previous estimates. The majority of the maximum and minimum temperatures are within ± 3 C of previous estimates. Changes in mean annual precipitation are greatest in coastal BC with some areas drier and other areas wetter than the previous estimates. Histograms show that most points are within 5 mm of previous estimates. A set of annual variables were determined for the 16 zones and 217 variants of the BEC system. As expected from the above-mentioned spatial analysis, there are differences between the new and past
3 estimates. Histogram plots indicate most changes are within 2% of previous estimates, but there are a few outliers. The greatest differences are in precipitation. The solar radiation estimates have not been evaluated against station data. It was found that the data base does not encompass some of the headlands and islets on the coast of BC. A fix for this could be to duplicate the outmost cells of the data base along the coast to ensure all of BC is covered. In conclusion, the tests indicate that Climate performs as well as previous versions. It is assumed that the difference seen in the spatial comparisons result from an improvement in resolution brought about by the improvements in methodology and increased data that went into creating the new PRISM data base. Acknowledgements Faron Anslow, PCIC, for supplying the PRISM grids and a copy 1971-2 temperature and precipitation normals. Adrian Walton, BCFLNRO, supplied the BEC grid point file.
4 Table of Contents Summary 2 Acknowledgments 3 List of Figures 5 Introduction 6 Changes to ClimateBC 6 Methods 7 Results 8 Discussion 12 References 13 Appendices Appendix 1: Monthly temperature and precipitation for 1951-8 normals 562 Meteorological Service of Canada weather stations calculated with ClimateBC V4 and without bilinear interpolation or adjustment for elevation. 14 Appendix 2: Monthly temperature and precipitation for 36 Meteorological Service of Canada weather stations for 1951-8 for ClimateBC with and without downscaling. 18 Appendix 3: Monthly temperature and precipitation and annual variables calculated with ClimateBC and measured at 31 Meteorological Service of Canada weather stations for the 1971-2 normals. 22 Appendix 4: Comparison of 1951-8 normals of monthly temperature and precipitation from ClimateBC and ClimateBC V4 with values for Meteorological Service of Canada weather stations. 27 Appendix 5: Comparisons of 1961-9 normals of annual variables from ClimateBC and ClimateWNA with values for Meteorological Service of Canada weather stations. 31 Appendix 6: Histograms of differences between ClimateBC and ClimateBC V4/ClimateWNA for 4 variables for 1951-8 normals. 35 Appendix 7: Selected annual variables (1961-9 normals) for BEC zones determined with ClimateBC and their difference from values calculated by ClimateBC. 36
5 Figures Figure 1: Mean annual temperature ( C) and mean annual precipitation (mm) for 1971-2 normals measured at 31 Meteorological Service of Canada weather stations and calculated using ClimateBC and ClimateWNA. 9 Figure 2: Summer heat:moisture index and annual reference evaporation (mm) for 1971-2 normals calculated for 31 Meteorological Service of Canada weather stations, and calculated for ClimateBC and ClimateWNA. 9 Figure 3: Difference between ClimateBC and V4 for mean annual temperature ( C) and mean annual precipitation (mm). 1 Figure 4: Difference between ClimateBC and V4 for July maximum temperature ( C) and January minimum temperature ( C). 11 Figure 5: Histograms of difference in mean annual temperature ( C) and percentage difference in mean annual precipitation (mm) between ClimateBC and ClimateWNA for the 217 BEC variants. 12 Figure 6: Histograms of difference in mean temperature of the coldest month (MCMT C) (left panel) and percentage difference in summer heat: moisture index (SHM) (right panel) between ClimateBC and ClimateWNA for the 217 BEC variants. 12
6 Introduction ClimateBC is an update to ClimateBC V4 and the BC component of ClimateWNA. ClimateBC is a software page that produces high spatial resolution estimates of temperature and precipitation and a number of derived climate variables for historic and future climates. Values for individual points are obtained by interpolation based on latitude, longitude and elevation of a base layer of temperature and precipitation normals (PRISM, Daly etal. 28). Each grid point at the centre of a PRISM tile has an elevation and a monthly value of temperature and precipitation. The interpolation is done by bilinear smoothing between the grid points and then adjusted to the elevation of a location using lapse rates (Wang et al. 212). Time periods other than the base period are obtained by the delta method using grids of change in temperature and precipitation from the base period (Climatic Research Unit Time Series, version 2.1 (CRU TS 2.1), Mitchell and Jones 25, Wang et al. 212). This document evaluates ClimateBC ability to produce climate normals and various derived variables for Meteorological Service of Canada weather stations. ClimateBC is also compared to estimates from ClimateBC and the BC component of ClimateWNA for the weather stations and for variants of the Biogeoclimatic Classification Systems. This testing is evaluation the accuracy of the PRISM base layer in combination with the downscaling methodology and the CRU anomaly data used to adjust to different time period than the base period. Changes to ClimateBC PRISM base layers: The original PRISM, created in 22, has base layers of the 1961-199 normals of monthly temperature and precipitation at 4 km resolution. The new PRISM layers are at 8 m resolution for 1971-2 period. Improved methodology (Daly et al. 28) and information from a substantially more weather stations in BC were used to generate the base layers. The work was done by Faron Anslow of the Pacific Climate Impacts Consortium in conjunction with Chris Daly s Climate Analysis Group at Oregon State University. Downscaling temperature to individual locations: Previous versions of the software obtained temperature lapse rates from a BC-wide equation of lapse rates as a function of latitude, longitude and elevation (Wang et al 26, 212). ClimateBC uses local lapse rates calculated dynamically based on temperature elevation of grid cells surrounding the point. Downscaling precipitation to individual locations: Previous versions of the software only had precipitation lapse rates that were inherent in the 4 km grid. No further downscaling was done. ClimateBC uses local lapse rates calculated dynamically based on precipitation and elevation of surrounding grid cells. Solar radiation: This variable was absent from previous versions of the software. ClimateBC contains a historic solar layer to allow monthly historic time series and adjustment of GCM predictions of solar radiation. The source of the historic solar layer is Hember (213).
7 CMIP5 projections of climate change: Previous versions of the software used the AR4 climate change projections and scenarios. ClimateBC includes projections from CMIP5. Also, projected values of solar radiation as well as temperature and precipitation are produced. Monthly time series for CMIP5 projections of climate change: Previous versions of the software only had climate projections for the 22s, 25s and 28s period normals. ClimateBC adds monthly time series of projections of temperature, precipitation and solar radiation for 214 to 21 for three projects from one GCM. Methods Comparison with measured data: ClimateBC calculation of climate normals are compared with those based on measurements at Meteorological Service of Canada weather stations. The normals periods were 1951-198 (the original test data set - Wang et al. 26, 212), 1961-199 (the PRISM base layers for previous versions of the software and reference period for the climate change data) and 1971-2 (the new PRISM base layers). Monthly and annual temperature and precipitation and derived variables such as degree days, frost free period, and evaporation, were generated with the software via an input file of station identification, latitude, longitude and elevation. The previous versions of the software used were ClimateBCV4 and ClimateWNA. The normals for the weather stations were obtained from Environment Canada. There are 562 stations for 1951-8 for ClimateBC V4 and 36 for ClimateBC (does not include adjacent areas in Alberta and the Yukon), 185 for 1961-9 (BC only) and 31 for 1971- (BC only). Comparison of original and new PRISM layers: This was done by running ClimateBC V4 and without any bilinear interpolation between tiles or downscaling to specific elevations. The old layers for 1961-9 period and new layers for 1971- where both adjusted using the CRU anomaly data base to the same period of 1951-8. The software was used to generate 1951-8 normals for 36 locations in the Meteorological Service of Canada weather stations. Spatial comparison: A BC-wide spatial comparison was made of ClimateBC and V4 for selected variables. Maps of the difference between ClimateBC and ClimateBC V4 were generated at 8 m grid for the 1961-9 period. The data were generated via an input file of station identification, latitude, longitude and elevation for each point. These data files were used to create files suitable for input to ArcGIS. Assessment of how the changes might influence climate descriptions: An extensive amount of work has been published using previous versions of the software. It is useful to see if the current versions produce data that might result in large differences in climatic description of areas of BC. This part of the analysis looks at changes to the estimates of climates within units of the Biogeoclimatic Classification System (zones and variants). Input to the software was a file with 6 points for each of the 217 BEC variants. The points are distributed such that they are at least 1 m from boundaries of other units. Solar radiation: This test consisted of running the software to ensure that reasonable values were generated. Testing of the solar data is being done by Hember (213).
8 CMIP5 projections: This test consisted of running the software to ensure that reasonable values were generated. There are no measured data or previous estimates from earlier versions of the software. Results Comparison of original and new PRISM layers: The software was used to generate 1951-8 normals for 36 stations without bilinear smoothing of downscaling. ClimateBC produced a marginal increase in R 2 over ClimateBC V4 for temperature. However, there was a slight decrease in some months for precipitation (Appendix 1). The reason for this is not known. It may be the finer resolution of the new PRISM layers has increased the variability. Downscaling of PRISM: Each grid point at the centre of a PRISM tile is given with an elevation and a monthly value of temperature and precipitation. The downscaling improves the estimate of the monthly temperature and precipitation (Appendix 2) over using the tile value. The improvements are small because of the small size of the tiles. The improvement was much great for the previous 4 km version of PRISM. In this version there was no improvement by downscaling of precipitation with lapse rates, only the bilinear smoothing was applied. Monthly and annual temperature and precipitation for 1971-2: This test evaluates the downscaling and interpolation methods of ClimateBC. There is excellent agreement in monthly temperature and precipitation between ClimateBC estimates and station data (Appendix 3). The R 2 values are high, the slope of the line is close to 1 and intercepts close to zero. The annual values and derived variables show similar excellent agreement between measured and ClimateBC values. ClimateBC provides a slight improvement over ClimateBC V4 for temperature, but produces slightly lower R 2 for precipitation. This is consistent with the results noted above comparing the original and new PRISM layes The earlier version of the software uses 1961-9 as the reference period and this requires and extra step of adjusting these data to a 1971-2 period using the CRU anomaly data. Figure 1 shows the estimates from ClimateBC and ClimateWNA V4 for mean annual temperature and mean annual precipitation. This can also be seen in the monthly data (not shown). However, as noted for the monthly data the estimation of precipitation is not as good for compared to (Figure 1). ClimateBC tends to underestimate precipitation, which may be a result of the underestimation of September and October precipitation noted above. This underestimation also affects the May to September precipitation total and thus the calculation of the summer heat moisture index (Figure 2). The improvement in estimating temperature can be seen in the temperature-based estimation of evaporation (Figure 2).
Station Station (mm) Station (C) Station (mm) 9 12 1 8 6 4 2 MAT 71- y =.9986x +.446 R² =.9533 y = 1.65x +.33 R² =.9786 Linear () Linear () 6 5 4 3 2 MAP 71- y = 1.571x - 49.76 R² =.979 y =.996x - 7.7156 R² =.9826 Linear () Linear () -4-2 2 4 6 8 1 12-2 -4, (C) 1 1 2 3 4 5 6, (mm) Figure 1: Mean annual temperature( MAT C) (left panel) and mean annual precipitation (MAP mm) (right panel) for 1971-2 normals measured at 31 Meteorological Service of Canada weather stations and calculated using ClimateBC (blue squares) and ClimateWNA (red squares). Regression lines generated with Excel. 25 SHM 71-9 Eref 71-2 15 1 y = 1.372x -.72 R² =.9717 y =.996x -.2898 R² =.9686 Linear () Linear () 8 7 6 5 4 3 y =.96x + 45.958 R² =.887 y =.9235x + 2.83 R² =.985 Linear () Linear () 5 2 1 5 1 15 2 25, 2 4 6 8 1, (mm) Figure 2: Summer heat:moisture index (SHM left panel) and annual reference evaporation (Eref mm) (right panel) for 1971-2 normals calculated for 31 Meteorological Service of Canada weather stations, and calculated for ClimateBC (blue squares) and ClimateWNA (red squares). Regression lines generated with Excel. Annual and monthly temperature and precipitation and derived variables 1951-8 and 1961-9: This test includes an evaluation of the adjustment of the 1971- reference period to other periods using the CRU anomalies. As with the 1971- normals, for 1951-8 ClimateBC shows a marginal improvement in monthly temperature (greater R 2 ) for the new PRISM, but a slight degradation for precipitation (Appendix 4) compared to ClimateBC V4. There appears to be a reduction in bias using the new PRISM layers with calculations closer to the 1:1 line and intercept closer to zero.
1 As with the 1951-9 analysis ClimateBC is a slight improvement over the previous software (in this case ClimateWNA ) for temperature in 1961-9 (Appendix 5). There is a minor degradation in performance for precipitation. Some outlier precipitation values, e.g., Glacier NP, were corrected. Derived variables such as heat:moisture indices, degree days and evaporation reflect the slight improvements over ClimateWNA. The 4 points well above the regression line for the frost free period are lighthouse stations (Appendix 5). This appears to be a result of the new PRISM having cooler temperatures near the ocean that the original PRISM (see next section). Spatial differences between ClimateBC and ClimateBC V4: It is expected that the software should predict weather station values reliably because these values went into generating the PRISM base layers. The strength of PRISM is its estimates of temperature and precipitation in areas far removed from points of measurement. Maps of BC for the difference between ClimateBC and ClimateBC V4 for selected variables were generated at 8 m grid. Figures 3 and 4 show that the new PRISM produces different climates in some compared to the original PRISM layers. In general, the new PRISM is warmer in the interior and cooler for the coast at high elevations, with differences in mean annual temperature ranging from -4.8 to +3.1. The cooler areas of northern BC likely result from ClimateBC V4 (and ClimateWNA ) using splined weather station data (Anusplin) for northeastern BC and northern Alberta rather than the original 4 km PRISM data. It was thought at the time that PRISM minimum temperature data for this area were too warm. This may not have been a valid assumption. Histogram plots show that on annual basis most of the points fall within ± 1 C of previous estimates. The majority of the maximum and minimum temperatures are within ± 3 C of previous estimates (Appendix 6). Figure 3: Difference between ClimateBC and V4 for mean annual temperature (MAT C) (left panel) and mean annual precipitation (MAP mm) (right panel). For MAT red indicates is warmer and blue is cooler than V4. For MAP blue indicates is wetter and brown drier than V4. Yellow and green indicates minor differences for MAT and MAP.
11 Figure 4: Difference between ClimateBC and V4 for July maximum temperature (Tmax7 C) (left panel) and January minimum temperature (Tmin1 C) (right panel). Figure red indicates is warmer and blue is cooler than V4.Yellow and green indicates minor differences. Changes in mean annual precipitation are greatest in coastal BC with some areas drier and other areas wetter than the previous estimates (Figure 3). Much of the interior has similar values for both versions of the software. The histograms (Appendix 6) show that most points for MAP are within 5 mm of previous estimates. Impact of changes on climates for BEC units: It is useful to have a comparison of past and current vesion in actual application of ClimateBC. Values for annual variables and differences between ClimateBC and ClimateWNA for zones are presented in Appendix 7. The differences mirror the tendencies shown by the maps in Figures 3 and 4. The greatest differences are in precipitation. Histogram plots of differences for the 217 BEC variants were generated for a selection of the annual variables. Figure 5 indicates that most of the differences in mean annual temperature of a variant are within ±1 C. For mean annual precipitation most changes are within 2% but there are a few outliers. The mean coldest month temperature (Figure 6) shows a similar distribution of differences to mean annual temperature. There are a few outliers up to 3.5 C different. As was noted earlier, this is a function of the use of splined data for northeastern BC that likely underestimated the minimum and average temperature. Most values for the summer heat:moisture index are within 15% of estimates from earlier versions of the software. Standard deviations on variant and zone means for the two versions of the software are similar.
Occurence Occurence Occurence Occurence 12 7 12 6 1 5 8 4 6 3 2 4 1 2-1 -.6 -.2.2.6 1 1.4 8 6 4-2 2 4 6 Difference in MAT (C) % Difference in MAP Figure 5: Histograms of difference in mean annual temperature (MAT C) (left panel) and percentage difference in mean annual precipitation (MAP mm) (right panel) between ClimateBC and ClimateWNA for the 217 BEC variants. 12 1 9 1 8 8 7 6 6 5 4 4 3 2 2 1-3.5-2.5-1.5 -.5.5 1.5 2.5 3.5-45 -35-25 -15-5 5 15 25 35 45 Difference in MCMT (C) % Difference in SHM Figure 6: Histograms of difference in mean temperature of the coldest month (MCMT C) (left panel) and percentage difference in summer heat: moisture index (SHM) (right panel) between ClimateBC and ClimateWNA for the 217 BEC variants. Solar radiation: The solar data provided does not fully cover BC. Certain coastal headland and islands are missing. This was seen in the inability to predict solar radiation for specific lighthouse stations. A fix for this could be to duplicate the outmost cells of the data base along the coast to ensure all of BC is covered. Discussion ClimateBC shows an improvement, though slight, over earlier versions in predicting measured temperatures at weather stations. There appears to be a slight reduction in the accuracy of the
13 precipitation estimates. Differences are seen between the new and earlier version in the spatial variation across BC and climatic descriptions of ecosystem climate. We believe the new PRISM layers better reflect the climates of these areas that the earlier versions; howver, we have no way to prove this. References Daly,C., M. Halbleib, J.I. Smith, W.P. Gibson, M.K. Doggett, G.H. Taylor, J. Curtis and P.P. Pasteris. 28. Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States. Int. J. Climatol. DOI: 1.12/joc.1688. Hember 213: Manuscript in reparation. Mitchell, T. D., and P. D. Jones. 25. An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol., 25, 693 712. Wang, T., A. Hamann, D.L. Spittlehouse and S.N. Aitken. 26. Development of scale-free climate data for western Canada for use in resource management. Int. J. Climatol. 26:383-397. Wang, T, A. Hamman, D.L. Spittlehouse and T.Q. Murdock. 212. ClimateWNA High-resolution spatial climate data for Western North America. J. Appl. Meteorol. Climatol. 51:16-29.
14 Appendix 1 Monthly temperature and precipitation for 562 Meteorological Service of Canada weather stations (1951-8 normals) for the 4 km grid of the original PRISM (red points) and 36 locations for the new PRISM at 8 m (black points). Bilinear interpolation between tiles or downscaling to the station elevation was not applied.
15 5 5 2 1 5 r 2 =.986 r 2 =.957 Tmax1( C) 1 5 5 1 5 5 r 2 =.983 r 2 =.942 Tmax2( C) 5 1 5 5 1 r 2 =.974 r 2 =.86 Tmax3( C) 5 1 15 5 1 15 r 2 =.962 r 2 =.75 Tmax4( C) 1 15 2 5 1 15 2 r 2 =.953 r 2 =.74 Tmax5( C) 1 15 2 25 1 15 2 25 r 2 =.954 r 2 =.733 Tmax6( C) 15 2 25 3 15 2 25 3 r 2 =.961 r 2 =.81 Tmax7( C) 15 2 25 15 2 25 r 2 =.959 r 2 =.82 Tmax8( C) 1 15 2 1 15 2 r 2 =.947 r 2 =.781 Tmax9( C) 2 4 6 8 1 14 2 4 6 8 1 14 r 2 =.964 r 2 =.855 Tmax1( C) 5 5 1 1 5 5 1 r 2 =.983 r 2 =.947 Tmax11( C) 15 1 5 5 15 5 5 r 2 =.983 r 2 =.948 Tmax12( C) Comparison between for Tmax
25 15 5 3 2 1 r 2 =.979 r 2 =.946 Tmin1( C) 2 15 1 5 25 15 5 5 r 2 =.974 r 2 =.947 Tmin2( C) 15 1 5 15 1 5 r 2 =.97 r 2 =.932 Tmin3( C) 8 6 4 2 2 4 5 5 r 2 =.941 r 2 =.881 Tmin4( C) 2 2 4 6 8 2 2 4 6 8 r 2 =.94 r 2 =.824 Tmin5( C) 2 4 6 8 1 12 2 4 6 8 1 12 r 2 =.874 r 2 =.779 Tmin6( C) 6 8 1 12 14 4 6 8 1 12 14 r 2 =.819 r 2 =.738 Tmin7( C) 4 6 8 1 12 14 2 4 6 8 1 12 14 r 2 =.838 r 2 =.777 Tmin8( C) 2 4 6 8 1 2 4 6 8 1 r 2 =.886 r 2 =.84 Tmin9( C) 4 2 2 4 6 6 2 2 4 6 8 r 2 =.935 r 2 =.884 Tmin1( C) 15 1 5 15 1 5 5 r 2 =.969 r 2 =.94 Tmin11( C) 25 15 5 25 15 5 5 r 2 =.976 r 2 =.938 Tmin12( C) Comparison between for Tmin
1 2 3 4 5 1 3 5 r 2 =.917 r 2 =.934 PPT1(mm) 1 2 3 4 1 2 3 4 r 2 =.942 r 2 =.959 PPT2(mm) 1 2 3 4 1 2 3 4 r 2 =.943 r 2 =.955 PPT3(mm) 1 2 3 5 15 25 35 r 2 =.932 r 2 =.943 PPT4(mm) 5 1 15 2 5 1 15 2 r 2 =.91 r 2 =.927 PPT5(mm) 5 1 15 2 5 1 15 r 2 =.852 r 2 =.846 PPT6(mm) 2 4 6 8 12 2 6 1 14 r 2 =.844 r 2 =.839 PPT7(mm) 5 1 15 5 1 15 2 r 2 =.88 r 2 =.85 PPT8(mm) 5 15 25 1 2 3 r 2 =.937 r 2 =.93 PPT9(mm) 1 3 5 2 4 6 r 2 =.952 r 2 =.95 PPT1(mm) 1 3 5 1 3 5 r 2 =.965 r 2 =.963 PPT11(mm) 1 2 3 4 5 2 4 6 r 2 =.948 r 2 =.957 PPT12(mm) Comparison between for PPT
18 Appendix 2 Monthly temperature and precipitation for 36 Meteorological Service of Canada weather stations for 1951-8 for the 8 m PRISM with no downscaling and with dynamical downscaling to the station elevation with ClimateBC (black points).
15 5 5 2 1 5 r 2 =.986 r 2 =.982 Tmax1( C) 1 5 5 1 5 5 r 2 =.984 r 2 =.974 Tmax2( C) 5 1 5 5 1 r 2 =.974 r 2 =.942 Tmax3( C) 5 1 15 5 1 15 r 2 =.962 r 2 =.891 Tmax4( C) 1 15 2 5 1 15 2 r 2 =.954 r 2 =.867 Tmax5( C) 1 15 2 25 1 15 2 25 r 2 =.954 r 2 =.884 Tmax6( C) 15 2 25 3 15 2 25 3 r 2 =.962 r 2 =.98 Tmax7( C) 15 2 25 15 2 25 r 2 =.959 r 2 =.93 Tmax8( C) 1 15 2 1 15 2 r 2 =.946 r 2 =.885 Tmax9( C) 2 4 6 8 1 14 2 4 6 8 1 14 r 2 =.964 r 2 =.914 Tmax1( C) 5 5 1 1 5 5 1 r 2 =.983 r 2 =.969 Tmax11( C) 15 1 5 5 15 5 5 r 2 =.983 r 2 =.978 Tmax12( C) Relationships between predicted and observed Tmax
25 15 5 3 2 1 r 2 =.979 r 2 =.977 Tmin1( C) 2 15 1 5 25 15 5 5 r 2 =.974 r 2 =.971 Tmin2( C) 15 1 5 15 1 5 r 2 =.97 r 2 =.963 Tmin3( C) 8 6 4 2 2 4 5 5 r 2 =.94 r 2 =.929 Tmin4( C) 2 2 4 6 8 2 2 4 6 8 r 2 =.94 r 2 =.889 Tmin5( C) 2 4 6 8 1 12 2 4 6 8 1 12 r 2 =.874 r 2 =.854 Tmin6( C) 6 8 1 12 14 4 6 8 1 12 14 r 2 =.82 r 2 =.87 Tmin7( C) 4 6 8 1 12 14 2 4 6 8 1 12 14 r 2 =.839 r 2 =.829 Tmin8( C) 2 4 6 8 1 2 4 6 8 1 r 2 =.887 r 2 =.881 Tmin9( C) 4 2 2 4 6 8 6 2 2 4 6 8 r 2 =.935 r 2 =.927 Tmin1( C) 15 1 5 5 15 1 5 5 r 2 =.969 r 2 =.964 Tmin11( C) 25 15 5 25 15 5 5 r 2 =.976 r 2 =.973 Tmin12( C) Relationships between predicted and observed Tmin
1 2 3 4 5 1 3 5 r 2 =.918 r 2 =.93 PPT1(mm) 1 2 3 4 1 2 3 4 r 2 =.943 r 2 =.936 PPT2(mm) 1 2 3 4 1 2 3 4 r 2 =.943 r 2 =.933 PPT3(mm) 1 2 3 5 15 25 35 r 2 =.931 r 2 =.94 PPT4(mm) 5 1 15 2 5 1 15 2 r 2 =.911 r 2 =.96 PPT5(mm) 5 1 15 2 5 1 15 r 2 =.851 r 2 =.851 PPT6(mm) 2 4 6 8 12 2 6 1 14 r 2 =.845 r 2 =.84 PPT7(mm) 5 1 15 5 1 15 2 r 2 =.879 r 2 =.879 PPT8(mm) 5 15 25 1 2 3 r 2 =.936 r 2 =.916 PPT9(mm) 1 3 5 2 4 6 r 2 =.953 r 2 =.942 PPT1(mm) 1 3 5 1 3 5 r 2 =.966 r 2 =.961 PPT11(mm) 1 2 3 4 5 2 4 6 r 2 =.948 r 2 =.944 PPT12(mm) Relationships between predicted and observed PPT
22 Appendix 3 Monthly temperature and precipitation and annual variables calculated with ClimateBC and measured at 31 Meteorological Service of Canada weather stations for the 1971-2 normals.
25 15 5 5 25 15 5 5 r 2 =.985 Tmax1( C) 2 1 5 1 2 1 5 1 r 2 =.983 Tmax2( C) 2 1 5 1 2 1 5 1 r 2 =.971 Tmax3( C) 1 5 1 15 1 5 1 r 2 =.961 Tmax4( C) 5 1 15 2 5 1 15 2 r 2 =.957 Tmax5( C) 1 15 2 25 1 15 2 25 r 2 =.961 Tmax6( C) 15 2 25 3 15 2 25 3 r 2 =.966 Tmax7( C) 1 15 2 25 3 1 15 2 25 3 r 2 =.958 Tmax8( C) 5 1 15 2 5 1 15 2 r 2 =.943 Tmax9( C) 5 5 1 15 5 5 1 15 r 2 =.971 Tmax1( C) 2 1 5 1 2 1 5 1 r 2 =.987 Tmax11( C) 2 1 5 2 1 5 r 2 =.986 Tmax12( C) Monthly maximum temperatures
3 2 1 3 2 1 r 2 =.977 Tmin1( C) 3 2 1 3 2 1 r 2 =.976 Tmin2( C) 3 2 1 5 3 2 1 5 r 2 =.969 Tmin3( C) 2 1 5 5 2 1 5 r 2 =.947 Tmin4( C) 5 5 1 5 5 1 r 2 =.922 Tmin5( C) 2 4 6 8 1 12 2 4 6 8 1 12 r 2 =.898 Tmin6( C) 4 6 8 1 12 14 4 6 8 1 12 14 r 2 =.869 Tmin7( C) 2 4 6 8 1 12 14 2 4 6 8 1 14 r 2 =.874 Tmin8( C) 5 1 5 1 r 2 =.93 Tmin9( C) 1 5 5 1 5 5 r 2 =.94 Tmin1( C) 25 15 5 5 25 15 5 5 r 2 =.972 Tmin11( C) 3 2 1 5 3 2 1 5 r 2 =.976 Tmin12( C) Monthly minimum temperatures
1 3 5 1 3 5 r 2 =.961 PPT1(mm) 1 2 3 4 1 2 3 4 r 2 =.966 PPT2(mm) 1 2 3 4 1 2 3 4 r 2 =.967 PPT3(mm) 1 2 3 1 2 3 r 2 =.969 PPT4(mm) 5 1 15 2 25 5 1 2 r 2 =.954 PPT5(mm) 5 1 15 2 5 1 15 2 r 2 =.945 PPT6(mm) 5 1 15 5 1 15 r 2 =.928 PPT7(mm) 5 1 15 2 5 1 15 2 r 2 =.935 PPT8(mm) 1 2 3 4 1 2 3 4 r 2 =.964 PPT9(mm) 2 4 6 2 4 6 r 2 =.969 PPT1(mm) 2 4 6 2 4 6 r 2 =.965 PPT11(mm) 2 4 6 2 4 6 r 2 =.966 PPT12(mm) Monthly precipitation
1 5 5 1 1 5 5 1 r 2 =.979 MAT( C) 8 1 14 18 22 8 1 14 18 22 r 2 =.942 MWMT( C) 3 2 1 3 2 1 r 2 =.984 MCMT( C) 1 2 3 4 1 2 3 4 r 2 =.984 TD( C) 1 3 5 1 3 5 r 2 =.971 MAP(mm) 2 6 1 2 6 1 r 2 =.962 MSP(mm) 2 4 6 2 4 6 r 2 =.969 AHM 5 1 15 2 5 1 15 2 r 2 =.97 SHM Calculated annual climate variables
27 Appendix 4 Comparison of 1951-8 normals of monthly temperature and precipitation from ClimateBC and ClimateBC V4 with values for Meteorological Service of Canada weather stations.
15 5 5 2 1 5 and v4 (red) outputs r 2 =.986 r 2 =.968 Tmax1( C) 1 5 5 1 5 5 and v4 (red) outputs r 2 =.983 r 2 =.976 Tmax2( C) 5 1 5 5 1 and v4 (red) outputs r 2 =.974 r 2 =.949 Tmax3( C) 5 1 15 5 1 15 and v4 (red) outputs r 2 =.962 r 2 =.924 Tmax4( C) 1 15 2 5 1 15 2 and v4 (red) outputs r 2 =.954 r 2 =.919 Tmax5( C) 1 15 2 25 1 15 2 25 and v4 (red) outputs r 2 =.954 r 2 =.917 Tmax6( C) 15 2 25 3 15 2 25 3 and v4 (red) outputs r 2 =.962 r 2 =.937 Tmax7( C) 15 2 25 15 2 25 and v4 (red) outputs r 2 =.959 r 2 =.932 Tmax8( C) 1 15 2 1 15 2 and v4 (red) outputs r 2 =.947 r 2 =.912 Tmax9( C) 2 4 6 8 1 14 2 4 6 8 1 14 and v4 (red) outputs r 2 =.964 r 2 =.949 Tmax1( C) 5 5 1 1 5 5 1 and v4 (red) outputs r 2 =.983 r 2 =.979 Tmax11( C) 15 1 5 5 15 5 5 and v4 (red) outputs r 2 =.983 r 2 =.965 Tmax12( C) Comparison between and v4 for Tmax
25 15 5 3 2 1 and v4 (red) outputs r 2 =.979 r 2 =.95 Tmin1( C) 2 15 1 5 25 15 5 5 and v4 (red) outputs r 2 =.974 r 2 =.96 Tmin2( C) 15 1 5 15 1 5 and v4 (red) outputs r 2 =.97 r 2 =.959 Tmin3( C) 8 6 4 2 2 4 5 5 and v4 (red) outputs r 2 =.94 r 2 =.933 Tmin4( C) 2 2 4 6 8 2 2 4 6 8 and v4 (red) outputs r 2 =.93 r 2 =.888 Tmin5( C) 2 4 6 8 1 12 2 4 6 8 1 12 and v4 (red) outputs r 2 =.873 r 2 =.861 Tmin6( C) 6 8 1 12 14 4 6 8 1 12 14 and v4 (red) outputs r 2 =.819 r 2 =.84 Tmin7( C) 4 6 8 1 12 14 2 4 6 8 1 12 14 and v4 (red) outputs r 2 =.838 r 2 =.827 Tmin8( C) 2 4 6 8 1 2 4 6 8 1 and v4 (red) outputs r 2 =.887 r 2 =.869 Tmin9( C) 4 2 2 4 6 8 6 2 2 4 6 8 and v4 (red) outputs r 2 =.935 r 2 =.912 Tmin1( C) 15 1 5 5 15 1 5 5 and v4 (red) outputs r 2 =.969 r 2 =.961 Tmin11( C) 25 15 5 25 15 5 5 and v4 (red) outputs r 2 =.976 r 2 =.949 Tmin12( C) Comparison between and v4 for Tmin
1 2 3 4 5 1 3 5 and v4 (red) outputs r 2 =.917 r 2 =.938 PPT1(mm) 1 2 3 4 1 2 3 4 and v4 (red) outputs r 2 =.942 r 2 =.964 PPT2(mm) 1 2 3 4 1 2 3 4 and v4 (red) outputs r 2 =.944 r 2 =.957 PPT3(mm) 1 2 3 5 15 25 35 and v4 (red) outputs r 2 =.933 r 2 =.941 PPT4(mm) 5 1 15 2 5 1 15 2 and v4 (red) outputs r 2 =.911 r 2 =.934 PPT5(mm) 5 1 15 2 5 1 15 and v4 (red) outputs r 2 =.853 r 2 =.868 PPT6(mm) 2 4 6 8 12 2 6 1 14 and v4 (red) outputs r 2 =.846 r 2 =.851 PPT7(mm) 5 1 15 5 1 15 2 and v4 (red) outputs r 2 =.882 r 2 =.871 PPT8(mm) 5 15 25 1 2 3 and v4 (red) outputs r 2 =.937 r 2 =.94 PPT9(mm) 1 3 5 2 4 6 and v4 (red) outputs r 2 =.953 r 2 =.955 PPT1(mm) 1 3 5 1 3 5 and v4 (red) outputs r 2 =.965 r 2 =.967 PPT11(mm) 1 2 3 4 5 2 4 6 and v4 (red) outputs r 2 =.948 r 2 =.963 PPT12(mm) Comparison between and v4 for PPT
31 Appendix 5 Comparisons of 1961-9 normals of annual variables from ClimateBC and ClimateWNA with values for Meteorological Service of Canada weather stations. Mean annual temperature ( C) Mean annual precipitation (mm) Mean warmest month temperature ( C) Mean coldest month temperature ( C) May to September precipitation (mm) Precipitation as snow (mm) Annual heat:moisture index Summer heat:moisture index Degree days below C (days) Degree days above 5 C (days) Degree days below 18 C (days) Degree days above 18 C (days) Frost free period (days) Beginning of frost free period (day of year) End of frost free period (day of year) Continentality ( C) Reference evaporation (mm) Climatic moisture deficit (mm) Other annual variables were not available for the weather stations.
Station (mm) Station (mm) Station (C) Station (C) Station (C) Station (mm) 12 1 8 MAT 61-9 y = 1.45x +.3664 R² =.9583 y =.9977x +.837 R² =.9814 45 4 35 3 MAP 61-9 y = 1.513x - 26.865 R² =.9659 y = 1.14x - 16.468 R² =.9791 6 4 Linear () Linear () 25 2 15 Linear () Linear () 2 1-2 2 4 6 8 1 12-2, (C) Mean annual temperature (C) 5 5 1 15 2 25 3 35 4, (mm) Mean annual precipitation (mm) 25 MWMT 61-9 MCMT 61-9 1 2 15 1 y = 1.1413x - 1.8618 R² =.9319 y =.9973x +.817 R² =.953 Linear () Linear () y = 1.115x +.343 R² =.9754 5 y = 1.28x +.1449 R² =.994-25 -2-15 -1-5 5 1-5 -1 Linear () Linear () 5-15 5 1 15 2 25, (C) Mean warmest month temperature (C) -2-25, (C) Mean coldest month temperature (C) 1 9 8 7 MSP 61-9 y = 1.361x + 3.6312 R² =.9599 y = 1.77x - 2.819 R² =.9715 16 14 12 PAS 61-9 y = 1.51x - 11.823 R² =.9191 y =.9161x + 1.3764 R² =.8583 6 5 4 3 Linear () Linear () 1 8 6 Linear () Linear () 2 4 1 2 2 4 6 8 1, (mm) May to September precipitation (mm) 2 4 6 8 1 12 14 16, (mm) Precipitation as snow (mm)
Station (days) Station (days) Station (days) Station (days) Station Station AHM 61-9 SHM 61-9 8 25 7 6 5 y = 1.785x -.661 R² =.9882 y =.9639x +.8153 R² =.9731 2 15 y = 1.126x - 3.74 R² =.9733 y =.9551x +.7892 R² =.964 4 3 Linear () Linear () 1 Linear () Linear () 2 5 1 1 2 3 4 5 6 7 8, 5 1 15 2 25, Annual heat:moisture index. Summer heat:moisture index DD< 61-9 DD>5 61-9 3 3 25 2 y = 1.383x - 72.786 R² =.9891 y = 1.53x - 18.24 R² =.9725 25 2 y = 1.56x - 64.466 R² =.8966 y =.9691x - 8.1374 R² =.9493 15 15 1 Linear () Linear () 1 Linear () Linear () 5 5-5 5 1 15 2 25 3, 472 (days) 5 1 15 2 25 3, 472 (days) Degree days less than C (days) Degree days above 5 C DD<18 61-9 DD>18 61-9 8 7 6 y = 1.86x - 41.563 R² =.9837 y = 1.172x - 183.69 R² =.9679 35 3 25 y = 1.4212x - 16.652 R² =.8217 y =.9667x - 7.182 R² =.8914 5 2 4 3 Linear () Linear () 15 1 Linear () Linear () 2 1 1 2 3 4 5 6 7 8, (days) Degree days less than 18 C (days) 5-5 5 1 15 2 25 3 35 4, (days) Degree days above 18 C (days)
Station (mm) Station (mm) Station (DoY) Station (C) Station (Days) Station (DoY) 35 FFP 61-9 25 bffp 61-9 3 25 2 y = 1.187x - 21.787 R² =.814 y = 1.498x - 8.1565 R² =.8752 2 15 y = 1.812x - 7.4642 R² =.7594 y = 1.32x - 2.741 R² =.8511 15 Linear () 1 Linear () 1 Linear () Linear () 5 5 5 1 15 2 25 3 5 1 15 2, (Days), (DoY) Frost free period (days) Start of frost free period (day of year) effp 61-9 TD 61-9 36 45 34 32 3 y = 1.157x - 47.128 R² =.8579 y = 1.854x - 24.9 R² =.892 4 35 3 y = 1.37x -.5827 R² =.9813 y = 1.383x -.7941 R² =.9887 28 26 Linear () Linear () 25 2 15 Linear () Linear () 24 1 22 5 2 2 22 24 26 28 3 32 34, (DoY) 1 2 3 4 5, (C) End of frost free period (day of year) Continentality (MWMT-MCMT) ( C) Eref 61-9 CMD 61-9 9 7 8 7 6 5 y =.949x + 1.78 R² =.9256 y =.9522x - 2.9776 R² =.9275 6 5 4 y =.9589x - 3.3545 R² =.9744 y =.9186x - 1.739 R² =.9671 4 3 Linear () Linear () 3 2 Linear () Linear () 2 1 1 2 4 6 8 1, (mm) -1 1 2 3 4 5 6 7, (mm) Reference evaporation (mm) Climatic moisture deficit (mm)
35 Appendix 6 Histograms of differences between ClimateBC and ClimateBC V4/ClimateWNA for 4 variables for 1951-8 normals. Variables are mean annual temperature (MAT C), mean annual precipitation (MAP mm), July maximum temperature (Tmax7 C) and January minimum temperature (Tmin1 C).
36 Appendix 7 Selected annual variables (1961-9 normals) for BEC zones determined with ClimateBC and their difference from values calculated by ClimateBC. Zonal averages are based on means for the variants within the zone. Table A5.1: 1961-8 normals for climate variables from ClimateBC MAT MAP MSP SHM DD5 PAS FFP EMT Eref CMD BAFA -2.7 993 437 21.8 357 589 51-46.3 293 18 BG 6.2 341 16 116.8 179 88 128-35.2 77 484 BWBS.1 667 338 42.2 9 277 89-45.2 424 96 CDF 9.5 1118 21 84.5 1986 48 211-17.1 655 262 CMA.1 3221 821 14.9 511 1972 63-39.5 374 16 CWH 6.1 2423 53 33.6 1337 478 152-25.8 541 83 ESSF.5 1172 39 31.5 656 688 7-41.7 432 72 ICH 3.2 896 338 46 1127 398 99-39.1 554 166 IDF 4.4 535 27 77.9 1313 198 15-37.2 622 343 IMA -1.1 1565 457 22.8 457 135 54-43.3 362 43 MH 3.1 3386 842 16.3 821 127 18-31.8 435 13 MS 2.4 768 25 55 917 383 8-39.4 531 225 PP 6.4 385 166 112.6 1726 13 129-34.2 724 482 SBPS 1.8 497 238 55.2 868 24 64-43.4 539 252 SBS 2.3 673 294 49.2 12 287 92-41 513 18 SWB -.9 14 422 3.4 615 528 77-44.6 342 4 Table A5.2: Difference between ClimateBC and ClimateWNA 4.72 MAT MAP MSP SHM DD5 PAS FFP EMT Eref CMD BAFA -.4 159-13 -2.5-76 182-4.8-21 -1 BG.7-21 -14 15.8 176-11 12 1.4 19 3 BWBS -.2 18 7 -.8-48 76-7.4 1 3 CDF -2-5 2.2 16 4 -.1 1 CMA -.3-49 -119.2-52 7 15.6-55 -6 CWH -118-41 3.5 18-9 2 -.4 18 ESSF -.1-18 -44 2.4-39 2 8 1.4-32 1 ICH.6-2 -24 5 88-14 1 1.9 5 23 IDF.4-12 -16 7.9 75-6 8 1 7 18 IMA -.6 5-7 1.3-122 96 8 2.1-61 -3 MH -.4 285 1.7-58 269 4-1.1-39 5 MS.2-21 -22 3.1 11-4 7.6-6 9 PP.6-12 -9 9.7 133-1 13 1.6 3 12 SBPS.2 33-12 2.6 33 28 6 -.1-1 9 SBS 6-21 2.9-1 21-2 -.2 4 19 SWB -.3 124-2 -1.8-13 18-1.8-12 5 MAT = Mean annual temperature ( C), MAP= Mean annual precipitation (mm), MSP= May to September precipitation (mm), SHM= Summer heat:moisture index, DD5 = Degree days above 5 C (days,. PAS = Precipitation as snow (mm), FFP = Frost free period (days), Eref = Reference evaporation (mm), CMD = Climatic moisture deficit (mm).
37 BAFA = Boreal Alti Fescu Alpine, BG = Bunch Grass, BWBS = Boreal White and Black Spruce, CDF = Coastal Douglas-fir, CMA = Coastal Mountain-heather Alpine, CWH = Coastal Western Hemlock, ESSF = Engelmann Spruce Sub-alpine Fir, ICH = Interior Cedar Hemlock, IDF = Interior Douglas-fir, IMA = Interior Mountain-heather Alpine, MH = Mountain Hemlock, MS = Montane Spruce, PP = Ponderosa Pine, SBPS = Sub-Boreal Spruce Pine, SBS = Sub-Boreal Spruce, SWB = Spruce Willow Birch.