No abstract

A procedure is described to construct time series of regional surface temperatures and is then applied to interior central California stations to test the hypothesis that century-scale trend differences between irrigated and nonirrigated regions may be identified. The procedure requires documentation of every point in time at which a discontinuity in a station record may have occurred through (a) the examination of metadata forms (e.g., station moves) and (b) simple statistical tests. From this "homogeneous segments" of temperature records for each station are defined. Biases are determined for each segment relative to all others through a method employing mathematical graph theory. The debiased segments are then merged, forming a complete regional time series. Time series of daily maximum and minimum temperatures for stations in the irrigated San Joaquin Valley (Valley) and nearby nonirrigated Sierra Nevada (Sierra) were generated for . Results show that twentieth-century Valley minimum temperatures are warming at a highly significant rate in all seasons, being greatest in summer and fall (Ͼ ϩ0.25°C decade Ϫ1). The Valley trend of annual mean temperatures is ϩ0.07°Ϯ 0.07°C decade Ϫ1. Sierra summer and fall minimum temperatures appear to be cooling, but at a less significant rate, while the trend of annual mean Sierra temperatures is an unremarkable Ϫ0.02°Ϯ 0.10°C decade Ϫ1. A working hypothesis is that the relative positive trends in Valley minus Sierra minima (Ͼ0.4°C decade Ϫ1 for summer and fall) are related to the altered surface environment brought about by the growth of irrigated agriculture, essentially changing a high-albedo desert into a darker, moister, vegetated plain.

[1] Temperature change of the lower troposphere (LT) in the tropics (20°S-20°N) during the period 1979-2004 is examined using 58 radiosonde (sonde) stations and the microwave-based satellite data sets of the University of Alabama in Huntsville (UAH v5.2) and Remote Sensing Systems (RSS v2.1). At the 29 stations that make both day and night observations, the average nighttime trend (+0.12 K decade À1 ) is 0.05 K decade À1 more positive than that for the daytime (+0.07 K decade À1 ) in the unadjusted observations, an unlikely physical possibility indicating adjustments are needed. At the 58 sites the UAH data indicate a trend of +0.08 K decade À1 , the RSS data, +0.15. When the largest discontinuities in the sondes are detected and removed through comparison with UAH data, the trend of day and night releases combined becomes +0.09, and using RSS data, +0.12. Relative to several data sets, the RSS data show a warming shift, broadly occurring in 1992, of between +0.07 K and +0.13 K. Because the shift occurs at the time NOAA-12 readings began to be merged into the satellite data stream and large NOAA-11 adjustments were applied, the discrepancy appears to be due to bias adjustment procedures. Several comparisons are consistent with a 26-year trend and error estimate for the UAH LT product for the full tropics of +0.05 ± 0.07, which is very likely less than the tropical surface trend of +0.13 K decade À1.

The techniques of nonlinear analysis are used to examine the behavior of the stable nocturnal boundary layer (SNBL) when it is subjected to changes in incoming radiation or in surface characteristics. A single‐column model and nonlinear bifurcation techniques are used to demonstrate that any atmospheric forcing, such as weak radiative forcing from greenhouse gases or cloud cover, can trigger a potentially significant positive feedback. Multiple solutions occur in some parameter spaces. This analysis shows that any forcing that decreases the stability, whether by increasing greenhouse gases or surface heat capacity, can cause large increases in surface temperature as the SNBL shifts from a weak turbulent regime, which allows the surface to cool, to a turbulent regime, which mixes warm air from aloft. Positive feedback may be a key factor in interpreting the long‐term observed nocturnal warming trend in the SNBL.