Study region
This study was focused on the top five tomato producing counties in California. According to the NASS (2017), top five tomato producing counties were Fresno, Kings, Merced, San Joaquin, and Yolo counties. These counties are located in central California from the southern end of the Sacramento Valley (38\(^\circ\)55′N) to the southern extent of the San Joaquin Valley (35\(^\circ\)56′N) (Fig. 1). Tomato production was about 3,982,000 tons for Fresno, 1,478,000 tons for Kings, 1,098,000 tons for Merced, 1,054,000 tons for San Joaquin, and 1,679,000 for Yolo counties in 2016, respectively. These counties make up 76% of the 2017 total contracted planted acreage for California (NASS 2017), and were selected for this study not only because they make up significant portion of processing tomato production in California but also represent a broad but typical production area for processing tomatoes.
Tomato transplanting dates
Nearly all commercial processing tomato fields in California are transplanted. Transplants are mechanically planted into fields starting from early March and continuing until early June (Hartz et al. 2008). To cover the processing tomato planting window, five planting dates were selected for input in the growing degree days (GDD) model: March 15, April 1, April 15, May 1, and May 15. Based on the expert opinions of tomato researchers and Extension professionals at the University of California, these dates would encompass early, average, and late planting periods where processing tomatoes are produced in California.
Modeled data
For this study, local scale projections were obtained from the Australian Community Climate and Earth-System Simulator (ACCESS 1.0) Global Circulation Model (GCM). ACCESS 1.0 model has an atmospheric resolution of 1.875° by 1.25° in the horizontal with 38 vertical levels and an atmospheric top at approximately 40 km. The ocean model has 50 vertical levels and 1° horizontal resolution, increasing to 1/3° near the equator (Bi et al. 2013). This modeling group is part of the Coupled Model Intercomparison Project Phase 5 (CMIP5) and successfully completed historical simulations of global average surface air temperatures. Selection of ACCESS 1.0 for this study was based on it performed better compared to other GCMs in California. The California department of water resources (DWR 2015) utilized 3-step model selection procedure to identify a subset of the better GCMs for developing assessments and plans for California water resource issues. Based on their analysis 10 CMIP5 GCMs passed the collective screening process and overall the ACCESS 1.0 provided the best performance. Based on that finding, we utilized ACCESS 1.0 GCM for this study.
For future projection trajectory, Representative Concentration Pathways 4.5 (RCP 4.5) and Representative Concentration Pathways 8.5 (RCP 8.5) were selected. The RCP 4.5 was developed by the modeling team at the Pacific Northwest National Laboratory’s Joint Global Change Research Institute (JGCRI), in the United States. It is a stabilization scenario in which total radiative forcing is designed to increase by 4.5 watts/m2. and stabilized shortly after 2100, without overshooting the long-run radiative forcing target level (Clarke et al. 2007; Smith and Wigley 2006; Thomson et al. 2011; Wise et al. 2009). RCP 8.5 was developed by the International Institute for applied systems analysis (IIASA), Austria which was designed to increasing greenhouse gas emissions over time, representative of scenarios in the literature that lead to high greenhouse gas concentration levels (Riahi et al. 2007). In other words, RCP 4.5 would mimic low emission scenario whereas RCP 8.5 would mimic “business as usual” or high emission scenario. Model outputs utilized in this study were statistically downscaled daily maximum and minimum temperatures over historical period of 1950–2005 and projected values over the period of 2006 to 2080; as Representative Concentration Pathways (RCP) scenarios typically start at 2006 (Collin et al. 2013).
Data are being made available by collaborators that include Bureau of Reclamation, Climate Analytics Group, Climate Central, Lawrence Livermore National Laboratory, Santa Clara University, Scripps Institution of Oceanography, US Army Corps of Engineers, US Geological Survey, and National Center for Atmospheric Research. We acknowledge the modeling groups, the Program for Climate Model Diagnosis and Intercomparison and the World Climate Research Program’s (WCRP) Working Group on Coupled Modeling for their roles in making available the WCRP Coupled Model Intercomparison Project (CMIP) phase 3 multi-model dataset. The Office of Science, US Department of Energy, provides support of this dataset.
Growing degree days (GDD) phenology model
Crops require certain amount of accumulated heat in order to move from one growth stage to the other. This is oftern referred to as physiological time. Growing degree-day models, which are mathematical combinations of various temperature thresholds, calculate cumulative heat units to relate to physiological times of various crops. Zalom and Wilson (1999) conducted a study on predicting phenological events for processing tomatoes in California. Based on their experiment with 536 datasets collected from commercial fields in California over 4 years, they reported that a growing degree day model with base temperature of 10 °C and upper cutoff of 30 °C significant improved prediction of various growth stages of processing tomatoes. According to their model validation, there was a significant accuracy and robustness in predicting physiological maturity when 1214 degree-days accumulated after planting.
Daily GDDs for the length of tomato growing season, i.e. duration from transplanting of tomatoes in the field to maturity was calculated using the following GDD model.
$$\begin{gathered} GDD=\frac{{Tmax~+~Tmin}}{2} - ~Tbase;\quad ~if~\frac{{Tmax~+~Tmin}}{2}>~Tbase~\;and\;~\frac{{Tmax~+~Tmin}}{2}<~Tcutoff \hfill \\ GDD=Tcutoff - ~Tbase;\quad ~if~\frac{{Tmax~+~Tmin}}{2} \geqslant ~Tcutoff \hfill \\ GDD=0;\quad ~if~\frac{{Tmax~+~Tmin}}{2} \leqslant ~Tbase \hfill \\ \end{gathered}$$
(1)
where \(Tmax\) and \(Tmin\) are maximum and minimum daily temperatures, respectively. \(Tbase\) and \(Tcutoff\) are defined as base temperature and cutoff temperatures, respectively. It is a horizontal cut-off model, which assumes that the degree days accumulations above the cutoff limit do not count.
Modeling application to analyze trends in growing season length
Procedure to estimate length of the growing season through the use of growing degree-days model and historical and future climate scenarios generated from the ACCESS 1.0 GCM model, over historical and projected timeframe is described in following steps.
Step 1: Determine transplant dates. Typical tomato transplanting dates for California is described in “Tomato Transplanting Dates” section.
Step 2: Gather daily minimum and maximum temperature data for study time period. For this research, minimum and maximum temperature for historical period (1950–2005) and future period (2006–2080) was obtained from ACCESS1.0 GCM downscaled data. Please refer to “Modeled Data” section.
Step 3: Calculate daily GDD for each of five planting dates for each of five locations using conditions defined in Eq. 1
Step 4: Calculate cumulative GDDs for each of five planting dates for each year for each county
Step 5: Determine day of the year for each year when accumulated GDDs reach minimum required cumulative GDDs for maturity as defined in “Growing Degree Days Phenology Model” section.
Step 6: Estimate trends in days to maturity and length of the growing season (duration between planting and maturity) for each of five planting dates for each of five counties through linear trend. Statistical significance was tested using the t-test (p < 0.05).