Bonfils et al. 10.1073/pnas.0700144104. |
Fig. 7. Maps of 1915-1979 d(t) trends in June-August maximum temperature computed lands, computed fomm the four datasets (in °C·decade-1), of the reference region (green outline) and of curent irrigation fraction >50% (black outline). All maps indicate a large daytime cooling concentrated in highly irrigated regions of the Central Valley.
Fig. 8. Maximum number of stations used to create the CRU2.1 (dot) and UW (plus) data sets. Only stations located below 500 m of elevation are indicated. The light gray shaded area outlines all irrigated areas (CIF > 0.1%), whereas the dark gray shaded area outlines lands that are the most irrigated (CIF > 50%).
Table 2. Distribution of irrigated areas in the studied region [focused on the Central Valley (CV)] and in the entire state of California (in percentage of total grid cells)
Mask | 0% | 0.1-10% | 10-20% | 20-30% | 30-40% | 40-50% | 50-60% | 60-70% | 70-80% | 80-90% | 90-100% | ||
CV | UW | 45.2 | 11.2 | 11.5 | 5.0 | 3.9 | 2.8 | 2.8 | 2.5 | 3.6 | 5.5 | 5.0 | 0.9 |
PRISM | 47.2 | 13.6 | 11.1 | 4.2 | 2. 6 | 2.2 | 2.1 | 2.1 | 3.2 | 5.1 | 5.1 | 1.5 | |
CRUs | 51.0 | 5.2 | 11.5 | 8.3 | 1.0 | 6.3 | 6.3 | 2.1 | 2.1 | 6.3 | 0.0 | 0.0 | |
CA | UW | 0.0 | 53.0 | 23.0 | 6.9 | 3.6 | 2.3 | 1.9 | 1.3 | 1.9 | 2.8 | 2.6 | 0.8 |
PRISM | 0.0 | 61.0 | 19.4 | 4.7 | 2.6 | 1.7 | 1.4 | 1.2 | 1.8 | 2.6 | 2.6 | 1.1 | |
CRUs | 0.0 | 33.3 | 36.1 | 13.9 | 2.2 | 4.4 | 3.3 | 2.2 | 1.1 | 3.3 | 0.0 | 0.0 |
For the studied region, the masked areas represent the regions above 500 m of elevation and the class (0.1-10%) represents the region of reference.
SI Text
Observational Data Sets.
The gridded irrigation data set was provided by the high-resolution (5' ´ 5') digital map of current irrigation fraction, which is based on the geographical distribution of areas equipped for irrigation around 1995 (1). Whether an equipped area is intensively irrigated or not irrigated remains part of the unmeasured uncertainties. The percent area covered by each CIF class in the studied region and in the entire state of California is indicated in SI Table 2.
The time series of irrigated land in the state of California since 1889 was linearly interpolated from paper records archived in U.S. Department of Agriculture censuses. Data since 1978 are available at www.nass.usda.gov/census/. Data from 1889 and prior 1978 were digitized from paper copies. Because annual variations in irrigated land cover are not available, it is more satisfactory to compare the reconstructed irrigation growth to a low-pass filtered temperature time-series (Fig. 2).
Region of Reference.
Using low-irrigated (where current irrigated fraction CIF ranges between 0.1% and 10%) as opposed to nonirrigated lands to define the reference time series is motivated by their greater proximity to intensively irrigated lands. Nonirrigated subregion would otherwise gather the largest urban areas, as well as remote regions likely influenced by different forcings and exposed to a wide variety of climate types. The reference region is plotted in SI Fig. 7. In all data sets, the reference region is composed of some coastal grid points but also of inland and Sierra foothills grid points. Masking data above 500 m of elevation regridded at the CRU, PRISM, or UW resolution does not critically influence the region of reference.
Calculation of Confidence Intervals for Linear Trends.
The 2s confidence intervals for the slope parameter of the estimated least-squares linear trend are adjusted for temporal autocorrelation (2). Based on the fact that least square linear regression residuals of a temperature time-series are generally not statistically independent, this adjustment assumes a lag-1 autocorrelation structure of the trend residuals, e(t). The lag-1 autocorrelation coefficient of e(t) is used to reduce the sample size, to inflate the standard error of the slope. This makes the test more difficult to reject the null hypothesis of a zero trend.
1. Döll P, Siebert S (2000), ICID J 49:55-66.
2. Santer BD, Wigley TML, Boyle JS, Gaffen DJ, Hnilo JJ, Nychka D, Parker DE, Taylor KE (2000) J Geophys Res 105:7337-7356.