Use of CLIMEX, Land use and Topography to Refine Areas Suitable for Date Palm Cultivation in Spain under Climate Change Scenarios

Global climate change resulting from changes in air and sea surface temperatures [1]; precipitation patterns, ocean level alterations [2] and ocean salinity [3,4] can affect agricultural productivity [5-7]. A review of literature reveals there have been many studies on the effect of climate change on ecological plant processes, such as plant growth [8], crop yield [9] and plant community interactions [3,10-14]. While McDermott [15] documented a greater than 55% crop failure in agricultural regions in India as a consequence of climate change, a comparable study reports that Sudan experienced a 50% crop failure in 2009 as a consequence of small changes in the quantity and pattern of precipitation, resulting from climate change [16]. Further, the potential effect on maize production of climate change has been reported [17] with indications of a 10% reduction by 2055, resulting in a loss of $2 billion per year in Africa and Latin America [17]. Shabani et al. [18] have reported on the effect of altered climate on global date palm production, in terms of long-term broad-scale shifts in the areas conducive to date palm cultivation. The total annual income from date palms in the Middle Eastern countries decreased from 1990 to 2000, due to water shortage and plant diseases resulting from climate change [19]; while an earlier example indicates that climate change caused a $438 million loss in wheat, a $116 million loss in grapes and a $67 million loss in sugar production in Australia and North America [14]. Substantial challenges for development and food security are posed by such losses in production and necessitate modeling the effects of climate change on agricultural production, especially crops important economically, to study the probable impact on agricultural output and distribution.

Date palm, which includes approximately 400 different species within the Arecaceae family, is a crop of economic importance in Middle Eastern countries, including Saudi Arabia, Egypt, Iran, Iraq, as well as some European countries, such as Turkey and Spain [20-24]. Date palm, being suited to a Mediterranean climate with salty and alkaline soils, has been grown in Spain for over 150 years and has become one of the main income sources for local farmers and the Spanish government [25]. Elche, located on the east coast of Spain, is one of Europe’s largest date palm plantations [26,27]. A statistical report indicates that date palm production in Spain reached 3,732 metric tons in 2003 [28]. Climatic parameters, plant diseases resulting from climate change and water shortage are factors that will impact on future date palm yields. Thus identifying how different regions of Spain may be affected by climate change is vital in terms of the future of date palm production in Spain.

Global climate models (GCMs) are important analytical tools in such disciplines as biogeography, evolution, ecology, conservation and invasive species management [29] and extensive use has been made of GCMs in biogeography, conservation biology and environmental management studies since 1996 [3,8,18,30-34]. The major function of climatic models is to illustrate plant sensitivity to climate change, through ascertaining relationships between plant and atmospheric sciences [3]. Such models use the current climate range to identify climate tolerance limits of particular species. Then, by modeling climate change scenarios, the alteration of the areas suitable for growth of vulnerable species can be projected into the future [35]. Such projections are valuable for the development of agricultural organizations’ long-term management strategies for sustainable production of economically important crops such as date palm (Phoenix dactylifera L.) under predicted future climate conditions [18].

There have been many models used to address the climate change factors, including bioclimatic models, ecological niche models (ENMs) [36], species distribution models (SDMs) [36-38] and CLIMEX [39]. CLIMEX is a computer-based software, designed for matching climates in ecology and has been widely used in estimating potential geographical distribution of different species in the future [11,40-44]. With CLIMEX, users can identify areas where selected species could establish and maintain or develop, based on predicted climate alterations. Through comparisons with other correlative modeling methods [45], CLIMEX has been evaluated as the most successful climate modeling software, for describing species’ responses to climate, due to its geographic range and powerful phonological observations [18,40,46,47].

Criticism of CLIMEX indicates that it does not include biotic interactions and dispersal in the modeling process [10,13]. Thus CLIMEX modeling output is based on limited climatic variables, such as overall maximum and minimum monthly temperatures (Tmax and Tmin), overall monthly precipitation levels (Ptotal) and the relative humidity taken at 09:00 h (RH09:00) and 15:00 h (RH15:00) [46,48]. This means that CLIMEX outputs are only based on climate predictions and the projected areas may still be unsuitable due to variables such as soil types, high slopes and urban areas. Thus, refining date palm suitability regions modeled with CLIMEX by investigating slopes and land use would result in a more accurate identification of the areas available for future date palm cultivation. Such refining can be undertaken using geographic information systems (GIS), remote sensing software (RS), species distribution models (SDMs) and ecological niche models (ENMs) [18,30,49-53]. In this study, it has been proposed that the distribution of P. dactylifera may effectively change as a result of climate alteration. With this possible outcome in mind, this research made use of the CLIMEX software package to determine the potential distribution of P. dactylifera under current and future climate and the location of the suitable slopes and land uses to assess the sensitivity of its distribution to climate change, and to evaluate the concurrent implications for cultivation of date palms for 2030, 2050, 2070 and 2100 in Spain.

Distribution of date palms (P. dactylifera)

Landsat images, with 30 m image resolution [54], the Global Biodiversity Information Facility (GBIF) [55], Missouri Botanical Gardens’ database [56] and other date palm literatures in CAB Abstracts databases [57] were used to collect data on P. dactylifera distribution and this information was supplemented by other date palm literature [22,24,58-70]. The GBIF database for Northern African countries, Saudi Arabia, Oman, Iraq and Iran had 314 records for P. dactylifera, but 64 records did not have geographical coordinates and were therefore removed, leaving 250 records. Some of these records were duplicates and were also deleted. In the end, 147 records from the GBIF database, 33 records from Missouri Botanical Gardens’ database and 43 records obtained from literature research were used to validate parameters in CLIMEX.

CLIMEX Software, climate data, climate models and climate scenarios

As one of the most comprehensive inferential modeling software packages, CLIMEX allows users to project the climatic impact on potential distribution of different species, based primarily on their current distribution [71]. The program makes use of known associations among climate, species distribution and their biological responses to project their future distributions [72]. Experimental observations of species’ growth response to temperature and soil moisture can be taken into account in CLIMEX [8,30,73,74]. The Ecoclimatic Index (EI) in CLIMEX is an average yearly index of climatic suitability based on regular weekly calculations of growth and stress indices. EI values range from 0 to 100, and a species can be established in an area if EI>0. EI values ranging from 0 to 10 indicate marginal potential habitats, while EI values >20 are highly favorable predictors [40]. Thus, an area with an EI>20 represents a potential area of high suitability from a climatic perspective, for introducing a particular species.

A variety of climatic parameter averages recorded between 1950 and 2000 are include in the CliMond 10’ gridded climate records and these were used to model the current distribution of date palm [46]. The meteorological database’s climatic parameters include overall maximum and minimum monthly temperature (Tmax and Tmin), overall monthly precipitation (Ptotal) and relative humidity as recorded at 09:00 h (RH09:00) and 15:00 h (RH15:00). These parameters were also used to predict climates for the mid- to late 21st century. The potential distribution of date palms under future climate was modeled using two Global Climate Models (GCMs), namely, CSIRO-Mk3.0 and MIROC-H (Center for Climate Research, Japan), with the A2 SRES scenario. These two GCMs were part of the CliMond dataset and were selected from 23 GCMs based on criteria such as the availability of required variables (including temperature, precipitation, sea level pressure and humidity for CLIMEX), small horizontal grid spacing in both GCMs and better representation of observed climate at local scales, compared with other GCMs [2,75,76].

The A2 SRES scenario was chosen as a possible climate scenario principally because of its inclusion of demographic and financial factors, as well as incorporating technological forces driving GHG emissions. The A2 scenario assumptions take into account factors of independent and self-reliant nations, such as population increases and regional economic development [77]. Future assumptions inherent in A2 SRES include high population growth, coupled with slow economic growth and extensive technological change. The A2 scenario assumes less extreme range of GHG emissions than scenarios such as A1F1, A1B, B2, A1T and B1 [77,78].

Model framing

The global distribution of date palm and the suitable areas for date palm cultivation at a global scale have been recently reported by Shabani et al. [18]. The reported CLIMEX parameters were used for Northern African countries, Saudi Arabia, Oman, Iraq, and Iran as the validation areas. Here, temperature and moisture response parameter values were obtained from the date palm physiology and growth model literature [16,22,35,58,67,68,79,80] and transformed into CLIMEX-compatible Temperature and Moisture Index parameters and cold, heat, dry and wet stress threshold values. The final values used for climatic parameters and growth indices in CLIMEX were 0.9 for the wet stress threshold (SMWS), 0.022 week–1 for the accumulation rate of wet stress (HWS), 4°C for the cold stress temperature threshold (TTCS), –0.01 week–1 for the frost stress accumulation rate (THCS), 46°C for the heat stress parameter (TTHS), 0.9 week–1 for the heat stress accumulation rate (THHS), 14°C for the limiting low temperature (DV0), 20°C for the lower temperature (DV1), 39°C for upper optimal temperatures (DV2), 46°C for the limiting high temperature (DV3), 0.007 for the limiting low moisture threshold (SM0), 0.013 for the optimal soil moisture (SM1), 0.81 for the upper optimal soil moisture thresholds (SM2) and 0.9 for the limiting high soil moisture (SM3) [24,25,55,58,62,65,70]. The data sets were output from CLIMEX and imported into ArcGIS (GIS) software for further processing and mapping.

Refining the CLIMEX outputs by suitable land uses and slopes

All vector data representing unsuitable land uses in Spain were taken from European Environment Agency Datasets [81]. The digital elevation model (DEM) of Spain on a regular 30 m grid from ASTER [82] (NASA, 2012) database were used to generate a slope surface, which was then reclassified into the two slope categories, mainly because overlaying the satellite images onto the date palm sites showed that 90% of the locations were on slopes less than 10 degree slopes (suitable 10 degree). Slope classes in raster format were converted to polygon shape files and queries were undertaken using attribute and location of unsuitable land uses including construction sites, coniferous forest, mineral extraction sites, rice fields, salt marshes, road and rail networks and associated land, green urban areas and bare rocks to extract the areas having suitable slopes and suitable land uses. The CLIMEX outputs for current time were overlaid on the location of suitable slopes and suitable land uses for the whole country and all those locations that satisfied the condition of EI>20 were selected.

Model validation under current climate

The distribution of P. dactylifera taken from different databases and an appropriate match between the Ecoclimatic Index (EI) from the CLIMEX model for Morocco, Algeria, Libya, Egypt, Saudi Arabia, Oman, Iraq and Iran is illustrated in Figure 1. The modeled results showed that southern Iran, Oman, Saudi Arabia, northern Egypt, northern Libya, Algeria and western Morocco have suitable climatic condition for P. dactylifera. While large parts of Libya and Algeria are modeled to have suitable climatic conditions for P. dactylifera in its current known distribution, inadequate distribution data were available from these regions. This could be due to a shortage of reporting from these areas as a result of lack of human distribution, or biotic factors such as competition or lack of dispersal opportunities could prevent this species from occurring in these areas [31,83-85]. Nearly 89% of the occurrence records fell within the suitable categories, confirming that the values selected for the various parameters in CLIMEX were optimum (Figure 1).

Figure 1: Current and potential distribution of P. dactylifera in a validation region based on the EI index.

Future climate

Figures 2 and 3 are illustrations of the Ecoclimatic Index for P. dactylifera, using the two global climate change models (CS and MR) and the A2 emission scenario, and show the projected distribution for the present, 2030, 2050, 2070 and 2100. The CS modeled results show that large areas of southern, western and central regions of Spain are predicted to become climatically more suitable for date palm cultivation and production between 2030 and 2100. The same model indicates that areas of northeast Spain may become suitable for P. dactyliferas’ cultivation between 2050 and 2100. The CS GCM predicts that northern and western Spain will become climatically unsuitable for date palm cultivation by 2100. This model shows roughly 11.5, 14, 19.5 and 28 million hectares will be conducive for date palm growth by 2030, 2050, 2070 and 2100 respectively (Table 1 and Figure 2). The CS model demonstrates that the areas located in 37.5° to 40°N and 1° to 5°W will become highly suitable for date palm cultivation from 2070 to 2100 while in this region the cold stress will remain and impose limitation for date growth until 2070 (Figures 2-6). Using the MR GCM, the model projects large areas in central and southwestern Spain will be climatically suitable for date palm growth by 2100. Additionally, the MR GCM results indicate that some areas in northern Spain may have an opportunity for date palm cultivation. As the climate condition is projected to be suitable by 2100. From the MR GCM, it can be seen that northwestern Spain will be climatically unsuitable, similar to the current situation. This model shows that areas from 37.5° to 40° N and 1° to 5° W and 40° N to 43° N and 4° W to 7° W will become highly conducive for date palm cultivation by 2100 (Figure 3). The MR GCM shows 12.2, 16, 24 and 33.8 million ha of Spain may have suitable climatic conditions for date palm cultivation by 2030, 2050, 2070 and 2100, respectively.

Table 1: Table 1- Results of CS and MR GCMs and the refined outputs by the location of unsuitable land uses and area less than 10? slopes for date palm cultivation for 2030, 2050, 2070 and 2100.

Figure 2: The climate (EI) for P. dactylifera at present and projected using CLIMEX under the CSIRO-Mk3.0 GCM running the SRES A2 scenario and for 2030, 2050, 2070 and 2100 for Spain.

Figure 3: The climate (EI) for P. dactylifera at present and projected using CLIMEX under the MIROC-H GCM running the SRES A2 scenario and for 2030, 2050, 2070 and 2100 for Spain.

Figure 4: The refined suitable areas for P. dactylifera under the CSIRO-Mk3.0 GCM running the SRES A2 scenario for 2030, 2050, 2070 and 2100 by the location of suitable land use and less than 10̊ slopes.

Figure 5: The refined suitable areas for P. dactylifera under the MIROC-H GCM running the SRES A2 scenario for 2030, 2050, 2070 and 2100 by the location of suitable land use and less than 10̊ slopes.

Figure 6: Location of cold and wet stresses in Spain for date palm cultivation for the present time and 2100.

Generally, both CSIRO-Mk3.0 and MIROC-H GCMs project similar trends for Spain. However, some differences can be seen in the projections. These differing results are due to the different predictions of future climate by the two GCMs (Figures 2 and 3 and Table 1) [86,87].

Future climate + Refined results

The refined CS outputs by areas less than 10°slope and the location of suitable land use for date palm cultivation shows that only about 7.38, 8.80, 12.81 and 18.72 million ha will be highly suitable for date plantation by 2030, 2050, 2070 and 2100 in Spain. The refined MR results show that about 7.87, 10.34, 15.14 and 22.86 million ha will be highly conducive for this species growth (Table 1).

Species distribution modeling was used as a tool with the aim of illustrating new conducive regions for date palm cultivation in Spain in the future. The results showed that, under current climate, large parts of eastern and western Spain are highly conducive to date palm cultivation (Figure 2), as there are no cold and wet stresses in those regions under the current climate (Figure 6).

There was some agreement in the projections of suitable areas for date palm cultivation between the CS and MR GCMs for Spain. Both MR and CS GCMs project that large parts of Spain will become more climatically suitable for date palm cultivation towards 2100 as a result of significant reduction in areas where currently cold and wet stresses are the major factors restraining date palm distribution (Figures 2, 3 and 6). For example, both models project that the areas with the highest potential for the establishment of date palm by 2100 are central Spain because cold stress may shift northwards (Figure 6). Additionally, both models projected that this unsuitability may be limited to the northern and northwestern regions of Spain due to significant increases in wet stress (Figure 6).

A comparison of the MR and CS models shows some projection differences. While the MR GCM projected that most regions in central Spain could become climatically suitable for date palm growth, the CS GCM projections showed similar results but with less extreme change and effects of change, especially from 2070 to 2100. It should be noted that some differences in the results of CS and MR GCMs (Figures 2 and 3) are due to differences in future greenhouse emission patterns. In this context, the MR model predicts that temperatures will increase by approximately 4.31°C, while the CS model predicts an increase of only 2.11°C by 2100. While the CS model predicts a 14% decrease in future mean annual rainfall, the MR model predicts a mere 1% reduction [2,86,87].

Climate change modeling results indicate the probable changes in the potential distribution of P. dactylifera in the future. As climate changes, certain areas where P. dactylifera currently occurs may become climatically unsuitable, and the economies of those areas will consequently be affected. The results of this study thus have broad implications for future date palm production in Spain for local farmers and the Spanish government, in terms of the predicted increasing possibilities for cultivating date palm. Predictions suggest there will be more areas available that are unaffected by wet or cold stresses, and thus be more conducive for date palm cultivation. In other words, Spain will be able to cultivate these crops to a larger extent in the future compared with the present.

The necessity of refining the CLIMEX results using the slope factor was due to the observation of high correlation between species distribution and abundance as functions of topography which directly effects moisture and fertility gradients [88,89]. Also, most of the studies on distribution of tree species have shown that topography is one of the most important factors affecting yields by causing problems in air permeability, hydraulic conductivity and root developments [90]. Furthermore, many studies have concluded that transportability of the sediment as a consequence of slope creates diverse microsite depending upon the soil type, geology and latitude [89,91,92].

A comparison between the mean CS GCM output and the mean refined results using slope and land use shows that only about 65% of the CS outputs will be suitable for date palm cultivation because of unsuitability in slopes and land uses. The slope map indicates that 34 million ha of Spain has less than 10° slopes and the Spain land use map showed that 2 million ha has highly unsuitable land use for date palm cultivation. For instance, wetland in Spain covers 1.1 million ha and this cannot be used as suitable land for date palm cultivation. From the MR results, refined using the suitable slope and land use, it can be seen that only 65% of the area which was projected to be climatically suitable may be conducive for date palm cultivation while the rest will be unsuitable due to unsuitable slopes and land uses.

A comparison of the results generated by the two models and then refined by slope and land use showed an average difference of about 15%. In the CLIMEX modeled results, the areas that are commonly projected by both GCMs was about 67%. Once the two modeled outputs were refined utilizing land use and slope data, results showed that 83% of areas projected by the two models overlapped for 2030, 2050, 2070 and 2100 (Figure 4 and 5).

It should be noted that there is a need to take into account additional non-climatic factors such as water access to the CLIMEX modeling output. For example, date palm can definitely cope with drought since it has evolved some strategies including high level of water absorption and widespread root system to protect itself in the face of stressors. However, in some areas date palm could only be cultivated (for fruit production) because it is irrigated. Thus, when there is no rainfall or irrigation, fruit production is greatly diminished. Therefore, water availability is an important non-climatic factor and could be a suitable refinement tool to refine the CLIMEX modeling results.

Such modeling allows us an insight into potential impacts of climate change on agricultural production and enables us to take appropriate steps to minimize negative impacts, as well as to take advantage of some scenarios that may arise in the future. Distribution maps such as those used to illustrate this study can assist in the formulation of production methods for cost-effective agricultural crops of the future. Valuable information on current and future changes of areas suitable for date palm cultivation can be deduced from maps such as those shown in Figures 4 and 5 and may encourage plantation owners and managers to consider the future implications of present day management decisions. The methods utilized in this study have equal application for other agricultural crops. Major factors that contributed to the success of this study were: (a) all variables needed for CLIMEX, including temperature, precipitation, sea level pressure and humidity, were readily available; (b) well-documented historical data on date palm distribution was incorporated; and (c) the popular GCMs chosen for this study were on account of (I) their small horizontal grid spacing and (II) their better representation of observed climate at the local scale, compared with GCMs not used in this study. The limitations of this study should be noted. For example, CLIMEX results are based on the response of a species to climate, and non-climatic parameters such as water availability, biotic interactions, soil type and competition between species have not been taken into account. Thus areas shown as being suitable are only based on future climate, suitable slopes and land uses. Large parts of these areas may be unsuitable due to the unsuitability of soil types. Future research could consider these factors in addition to CLIMEX-based modeling.