Climate indices

For the characterization and definition of production potential of Baja California, four indices of temperature which are described below were used.

Development degree days (GDD) is an index ofheat buildup, commonly used to define the suitability of a region for growing grapes (^{Diffenbaugh et al., 2011}; ^{Hall and Jones, 2014}). It is based on the observation that the vines begin their active growth in the spring when the average air temperature for five consecutive days reaches 10 °C (^{Amerine and Winkler, 1944}).

According to this method the GDD are calculated as shown below:

Where: Tm= mean daily temperature (°C).

In the present study, GDD were calculated from an image of average temperature for the period from April to October, when the growth cycle of the vine on the Costa de Ensenada occurs. Once calculated the GDD for different regions ofthe State, we proceeded to their classification based on the system of viticultural regions proposed by ^{Amerine and Winkler (1944)} , which considers five classes ofviticultural conditions based on the style and quality ofwines which can be produced in a given climate. In the climate I (<1 370 GDD) the region of early maturing varieties reach high quality. In region II (1 371-1 650 GDD) most ofthe early and intermediate varieties produce wines of good quality, light to medium body and good balance. The region III (1 651-1 930 GDD) has a climate suitable for high production of wines at good standard full-body dry to sweet dessert wines. The region IV (1 931-2 205 GDD) is favourable for high yields and sweet wines, and is generally lower for table wines. Region V (> 2 205 GDD) produces generally fill table wine or fortified wine, or is it better to varieties oftable grapes or raisins.

Frost-free period (PLH): to calculate the frost-free period, which gives us the days or months available for the growing season, we used the method proposed by ^{Snyder et al., (2005)} . The calculations were made considering a probability of 90% frost. As the minimum temperature used in the calculations of critical value was 0 °C, since at this temperature damage to plant shoots arise in the spring.

The following model was used to obtain the image of PLH:

Where: PLH= free of frost (days) Period; TIOA= average minimum temperature for the period from October to April (°C).

PLH data were grouped into the following categories: A) <135 days; B) 136-155 days; C) 156 to 175 days; D) 176 195 days and E)> 195 days frost-free.

Cold hours (CH): The deciduous fruit trees and vines require a rest period in the winter where low temperatures influence the percentages of sprouting in spring (^{Kliwer and Soleimani, 1972}; ^{Botelho et al., 2007}). Because no data are collected every hour, but once in the day, cold hours were calculated using the minimum daily average temperatures recorded at each station in the period from November to February, according to the method of Da Mota (1983), cited by Ortíz (1987) in which the following equation is used:

Where: CH= cold hours accumulated in the winter period; Tm= average temperature of the months of November to February.

In order to obtain the total of CH over the year subtotals, CH obtained for each of the four months were added. The CH were classified into four classes: 1) <50 CH; 2) 51-150 CH; 3) 1 51-250 CH and 4)> 250 CH.

Average maximum temperature for the period fromApril to October (TMM). For analysis ofpotential areas for growing grapes, it was considered as an important parameter values of maximum temperature of the period from April to October. We determined that the regions with temperatures above 32 °C were unsuitable for growing grapes, based on reports that said that the maximum temperature limit for metabolic processes of the vine is between 30 and 32 °C (^{Coombe, 1987}; ^{Gladstones, 2002}).

From climatic parameters matrices formed the basis ofdaily data, thematic images corresponding to each parameter were generated. The images were generated by raster format geographic interpolation process by IDRISI and Andes system (^{Eastman, 2006}). The methods included the application of a heat-high-lati model for temperature (^{Medina et al, 1998}). Thus themed images of maximum temperature and minimum monthly temperatures were obtained, from which some other seasonal or annual variables were obtained through processes of map algebra in the IDRISI system.

Identification of climate wine regions of Baja California

As a first step towards identifying the climate of wine regions in Baja California, a method of qualitative binary agro-ecological regionalization was used, i.e. two strata:

The parameters and strata considered are described in Table 1.

Table 1 Variables and criteria used to determine climate wine regions in Baja California. 

Based on the information just presented in the above table, the areas with less than 1 307 GDD and those with more than 2 205. This is excluded because the first value indicates the heat that builds up in the April-October period is insufficient to achieve fruit with enough sugar but with high acidity; while the second shows that the region accumulates so hot that the fruits can have high sugar content but low acidity, making them unfavourable for the production of table wines with good quality. The PLH parameter regions with less than 135 frost-free days, because this is a very short period to allow the growth of any vine cultivar were excluded; in other words the risk of frost damage to plants in these regions is quite high.

In CH were excluded regions with less than 150 cold hours as grape cultivars accumulate in the winter require at least that amount of cold. In the setting of maximum temperatures were removed regions with values above 32 °C, as this temperature limit proper photosynthetic activity of the vine, besides the risk of damage sustained by fruit stages of veraison to maturation. Finally, the areas occupied by urban centres and bodies of water were removed.

After excluding areas unsuitable for vine climatic parameters, we proceeded to perform a stratified environments based on the combination ofAHP and GDD. The strata considered for tríese two parameters are described in Table 2.

Table 2 PLH and GDD intervals to determine climate wine regions in B. C. 

The procedure was performed in Idrisi Andes system using map algebra module, running a multi-criteria reclassification processes involving imaging parameters and the superimposition of images reclassified analysis. Once the climatic regions were bound we proceeded to estimate the areas occupied by each of them by the AREA command system IDRISIAndes. The maps were generated with a level of detail of180 x 180 m (3.24 ha), and digital elevation map was used with a resolution of 90 x 90 m (0.81 ha). Finally, the maps were edited with conical Lambert projection through the ArcGIS system.

Climate change and the growing potential of Baja California

In order to simúlate the effects of climate change on the production potential ofthe regions identified, the information system ofclimate change (SICC) by INIFAP was used (^{Ruiz et al., 2010}); which is based on the use ofan assembly model and provides climate values weighted with 10 GCMs (mpi_ echam5, miub_echo_g, csiro_mk3_0, csiro_mk3_5, cccma_ cgcm3_1, giss_model_e_r, ncar_ccsm3_0, miroc3_2_hires, mri_cgcm2_3_2 ^a, ukmo_hadcm3). With that system were estimated GDD, PLH, CH and wine climatic regions for the decades 2011-2020, 2031-2040 and 2051-2060, considering the scenario ofgreenhouse gas emissions A2 (^{IPCC, 2007}). The management of this information is performed using Idrisi Andes system images in raster format.

Climate indices

GDD. In the Figure 1 are shown the cumulative GDD of the first day of April to October 30, calculated based on the method ^{Amerine and Winkler (1944)} . It is noted that in the present situation (period 1961-2010) in the State we can find the five wine regions described by these authors. As expected, the regions where less GDD accumulate are located in the highest parts ofthe State, as in the case ofthe Sierra de San Pedro Martir located west of San Felipe and Juárez east of Ojos Negros, where temperatures are lower. In such places the altitude above mean sea level exceeds 1 000 m. In contrast desert regions located at a much lower altitude and east ofthe State, have a large cumulative GDD (> 2 205 °C).

Figure 1 GDD (°C) accumulated in the period from April to October in the current conditions, and three future scenarios in Baja California. 

The analysis ofthe processed maps can envision the impact that climate change may have on the GDD in different regions of the State. It is noted that even through, the three future scenarios, the GDD increase statewide, with the result that for 2051-2060, only the higher altitude regions and some coastal areas of western State values less accumulate GDD to the 2205, the maximum suggested for good quality fruit for winemaking (^{Winkler et al., 1972}; ^{Gladstones, 2005}).

According to this, the wine valleys where currently such as Guadeloupe, St. Thomas, St. Vincent and Ojos Negros occur, after the year 2051 will be placed according to the classification of ^{Amerine and Winkler (1944)} , in the wine region V, that is, where the climatic conditions would not be favourable for quality wines.

PLH. The PLH map (Figure 2) shows the classification of the five regions of B .C., as the regions with higher altitudes above sea level, located in or near the mountains of San Pedro Martir and Juárez indicated by the white expected colour, have less than 135 days PLH, which limits its feasibility for growing grapes. By contrast the regions where the wine valleys are located, such as Guadalupe, Santo Tomas and San Vicente have PLH greater than 195 days, allowing the cultivation of the vine with minimal risk of frost damage.

Figure 2 PLH from November to February in the current conditions, and in three future scenarios in Baja California. 

CH. The data indicate that under current conditions and the three future scenarios, the major wine areas of the State have an accumulation of CH higher than 250 (Figure 3). Since most of the varieties of wine grapes require less than 250 CH for good bud break in the spring (^{Weaver and Iwasali, 1977}; ^{Botelho et al, 2007}), we can conclude that in general, this factor should not be a limiting factor for growing grapes in B. C.

Figure 3 CH accumulated in the period from November to February in the current conditions, and three future scenarios in B. C. 

TMM. In the 1961-2010 scenario, the TMM is observed in the period from April to October, regions with higher than 32 °C values, and identified in Figure 4 with intense green colour are located on the fringe of the desert-coastal east of the State, and include in the north to Mexicali Valley and Southern limit State. Due to its high temperatures, all that portion is discarded in this paper as an area with potential for growing grapes for wine. As shown in the following three scenarios, areas with TMM above 32 °C, increase with the passage of time, implying a significant reduction in areas with potential for growing grapes for wine in B. C. According to the 2051-2060 scenario the area between the west coast of the State and saws higher than 1 000 m (indicated by the light blue) heights, would retain the potential for the wine industry.

Figure 4 Mean maximum temperatures (TMM) in the period from April to October in the current conditions, and three future scenarios B. C. 

Climate change and growing potential

The interaction between climatic and bioclimatic variables identified 20 viticultural climatic regions in B. C., which are shown in Figure 5. In the 1961-2010 scenario producing regions of the best wines of Mexico such as the Valle de Guadalupe (Francisco Zarco), St. Thomas, and St. Vincent are located within growing regions III and IV of ^{Amerine and Winkler (1944)} with a cumulative GDD between January 1651 and 1930, and a frost-free period ofmore 195 days, which allows grape cultivars develop both short and long cycle. Also the West Valley Ojos Negros portion lies within this classification, while the eastern part is classified in the region II with less than 155 frost-free days. This means that only short cultivars intermediate cycle could be established in this part.

Figure 5 Stages of change of production potential of B.C., depending on climate change. 

In the 2011-2020 scenario, these regions retain the characteristics mentioned above, except for the eastern portion of San Vicente would happen to the wine region V, this is already not be suitable for yielding table wine quality. In the next stage (2031-2040), the Valley of Guadalupe, Santo Tomas and San Vicente Premium, would be classified in the V PLH wine region with more than 195, the above as a result of the increase in temperature. The Ojos Negros valley would retain the classification IV, except for the north would take the classification V.

Finally in the last scenario considered (2051-2060), the four valleys mentioned above would be classified in the region V with PLH more than 195 days. According to these latest predictions, within 37 years the current wine regions of Baja California where the best wines of Mexico, would cease to be to become unfit for quality table wines areas. Producing fruits for making table wines fill or fortified wines, as well as varieties of table grapes or raisins. This could change if new cultivars or clones that are adapted to the conditions of higher temperatures, new management practices are applied plants, the design of the vineyards is changed, or adjustments in the winemaking as has been suggested by some authors (^{Jones et al, 2005}; ^{Holland and Smit, 2010}; Duchéne et al., 2014; ^{Viguié et al, 2014}).

On the other hand, the analysis of the maps can identify areas under the effects of climate change could become the new producing quality wine regions in the State of B. C. The images show that in the long term (2051-2060), the regions around the mountains of Juárez and San Pedro Martir, as well as parts ofthe valleys of Ojos Negros and Trinidad have suitable conditions for growing grapes whose purpose is the development oftable wines. This condition would be for the coastal strip running from the south of the city of Ensenada to San Vicente.