Abstract

The impact of climate change on potato cultivation in Montenegro was assessed. Three scenarios (A1B, A1Bs and A2) for 2001–2030, 2071–2100 and 2071–2100, respectively, were generated by a regional climate model and compared with the baseline period 1961–1990. The results indicated an increase of temperature during the summer season from 1.3 to 4.8 °C in the mountain region and from 1 to 3.4 °C in the coastal zone. The precipitation decreased between 5 and 50% depending on the scenario, region and season. The changes in temperature and precipitation influenced phenology, yield and water needs. The impact was more pronounced in the coastal areas than in the mountain regions. The growing season was shortened 13.6, 22.9 and 29.7 days for A1B, A1Bs and A2, respectively. The increase of irrigation requirement was 4.0, 19.5 and 7.3 mm for A1B, A1Bs and A2, respectively. For the baseline conditions, yield reduction under rainfed cultivation was lower than 30%. For A1B, A1Bs and A2 scenarios, yield reductions were 31.0 ± 8.2, 36.3 ± 11.6 and 34.1 ± 10.9%, respectively. Possible adaptation measures include shifting of production to the mountain (colder) areas and irrigation application. Rainfed cultivation remains a viable solution when the anticipation of sowing is adopted.

INTRODUCTION

Montenegro (MNE) is a small Mediterranean country located in the Balkan Peninsula, South-Eastern Europe. The country is characterized by strong variation of climate and landscape features across its territory. The coastal area belongs to the central-southern part of the eastern Adriatic coast and it is characterised by a typical Mediterranean climate. The inlands have a temperate continental climate with four seasons, whereas higher zones in inlands have a cold mountainous climate. The average annual air temperature in the coastal area is around 14–15 °C. Yet, only 50–100 km into the inland area, there are locations with average air temperature below zero for more than 150 days per year (Rajković & Đurđević 2010). Therefore, the country is marked by strong variation of agroecological zones and other environmental factors (Knezevic et al. 2017).

The changes in climate could affect crop production in the coastal area, as well as Montenegro's inlands. The mean air temperature over the whole Balkan Peninsula has been rising in the last few decades, and the intensity of long dry spells has been increasing.

The shorter growing seasons for crops could be expected due to the increase in the air temperature (Čereković et al. 2010; Stričević et al. 2017). Many studies reported an increase of the crop water requirements and net irrigation requirements (NIR) in the Mediterranean region (Rodriguez-Diaz et al. 2007; Giannakopoulos et al. 2009), whereas some authors (Lovelli et al. 2010) foresee a decrease of NIR due to more favourable rainfall distribution. Therefore, some crops could benefit from warmer and longer growing seasons, whereas the others are expected to shift northwards, and to higher altitudes (Knezevic et al. 2017). The magnitude of changes may vary depending on latitude, temperature, water availability and agricultural practices among which irrigation and sowing dates are the most important.

Potato (Solanum tuberosum L.) is one of the most cultivated crops in the world and in Montenegro is grown in all agroecological regions, from coastal to high mountainous areas. The total cultivated land in 2014 was 2,137.7 hectares (1,645.1 ha of arable land and 492.6 ha of yards) (www.monstat.org) and the production was around 31.097 t. The yield of potato in Montenegro is very low compared with the other European countries due to low land fertility, inadequate agricultural management and rainfed production (Callaway et al. 2010; Jovovic et al. 2016).

The response of potato crop to climate change was assessed in several studies that focused on different regions and modelling approaches (Daccache et al. 2011; Franke et al. 2013; Raymundo et al. in press; Stričević et al. 2017). Most of the studies showed that yield would decline, although those studies considered models without atmospheric CO2 effect (Raymundo et al. in press). There is no change in potato yield according to a model that included atmospheric CO2 increase (Stockle et al. 2010). Supit et al. (2012) pointed out that potato production would shift to the cooler regions and an increase of atmospheric CO2 could mitigate the effect of the elevated temperature.

This study aimed to assess the impact of climate change on potato cultivation in Montenegro in terms of irrigation requirements and yield. The baseline period (1961–1990) was compared with three Special Report on Emissions Scenarios (SRES) generated by regional climate model EBU-POM (Eta Belgrade University – Princeton Ocean Model) (Rajković & Đurđević 2010). These scenarios were: i) A1B, referring to the period 2001–2030; ii) A1Bs, referring to the period 2071–2100, and iii) A2, referring to the period 2071–2100.

DATA AND METHODOLOGY

Study area

Montenegro (Figure 1) is located in a temperate zone with climate that varies between Mediterranean, sub-continental and continental over a relatively small area (13,810 km2). The relief is characterized by mountain chains intersected with narrow river valleys and coastal area. Very high altitudinal gradients made an impact on the local climate variability and agricultural production.

Figure 1

Geographic location of Montenegro (a) with the maps of administrative regions (b).

Figure 1

Geographic location of Montenegro (a) with the maps of administrative regions (b).

The first climatic zone comprises the Montenegrin coastal area and the Zetsko-Bjelopavlićka Plain, with surrounding hilly areas. This area is characterized by the Mediterranean and modified Mediterranean climate according to the Köppen criteria (Burić et al. 2014). The average annual temperature varies between 14 and 15 °C, whereas precipitation is abundant (1,300–2,500 mm) and occurred mainly during the autumn-winter-spring season, i.e. from October to April. As a result, summer months are hot (Csa) and exposed to long dry periods. The second climatic zone comprises the Karstic region in the continental part of Montenegro, up to 1,000 m altitude. This zone is characterized by a continental climate, comprising an etesian climate with a warm summer (Csb) and humid temperate climate with warm summer months (Cfb). The third climatic zone includes the pre-mountainous area (with altitude above 1,000 m) and the mountainous part of the country (with altitude above 1,500 m). This zone is situated in the north-western and eastern part of the country and has a humid cold temperate climate with warm summers (Dfb) and a humid boreal climate with cool summers (Dfc), respectively.

Agro-ecological zones and soil data

Montenegro is divided into five agro-ecological regions by means of climate, geology, soil productivity and agricultural production structure (Figure 1) (Knezevic et al. 2017). The coastal zone is situated along the Adriatic coast and covers about 12% of the territory. It is characterized by typical Mediterranean pedo-climatic conditions and crops. The Zetsko-Bjelopavlićki region accounts for the areas of Podgorica and Danilovgrad municipalities, and it comprises 14% of the total country territory. This is the main Montenegrin lowland region. Dominant soils are Cambisols, Fluvisols and Gleysols. This region has optimal conditions for agricultural production. The Karstic region is situated in the territory of Cetinje and Nikšić municipalities and covers 21% of the complete territory, but arable land makes up only up to 8% of the region. The dominant agricultural sectors are livestock production and beekeeping, whereas soils are shallow and scarce in terms of fertility.

The northern-mountainous zone is the largest region (32.5% of total area) and it includes all the municipalities of the central and northern parts of the country. This area is characterized by many plateaus, levelled fields, often with deeper soils, suitable for crop production. The Polimsko-Ibarski region covers 20.5% of the country area. This region has the largest share of arable land, 32.9% of the total arable land of Montenegro, or 62,000 ha in absolute terms. The prevailing soils are relatively fertile Fluvisols, Regosols and Cambisols, around the river valleys and on river terraces, and Dystric Cambisols on the moderately steep slopes. Data used in the study represent typical soil characteristics of five agro-ecological regions and they are presented in Table 1.

Table 1

Basic physical and chemical characteristics of typical soil groups dominating the area of Montenegro (Fuštić & Đuretić 2000)

Soil group Location Soil depth (cm) > 2 mm 0.02–0.002 < 0.002 2–0.02 (%) pH H2pH KCl CaCO3 (%) Organic matter (%) 
Colluvic Regosols Mrcevopoljea 0–30 10.3 40.7 9.6 49.7 7.09 6.39 4.5 2.99 
30–60 8.5 35.3 11.6 53.3 6.99 6.14 1.9 1.98 
70–100 3.9 42.3 15.4 42.3 7.34 6.54 4.3 1.64 
Rendzic Leptosols Previsa 0–30 19.4 39.1 17.8 43.1 7.7 6.7 13.7 9.1 
Mollic Leptosols Zagricab 0–10 52.6 29.7 6.4 63.9 7.6 6.65 13.7 17.03 
10–20 61 36.0 4.0 60.0 7.59 6.45 23.3 16.08 
Dystric Cambisols Kosuticac 1–14 27.9 43 13.1 43.9 4.97 4.11 0.0 6.08 
14–47 23.9 38.9 18.7 42.4 5.26 4.32 0.0 2.52 
50–70 25.2 39.2 23.1 37.8 5.21 4.36 0.0 2.22 
Eutric Cambisols M. Zdrebaonikd 0–26 43.1 8.4 48.5 6.72 5.26 0.0 3.75 
30–60 41.6 28.1 30.3 6.72 5.71 0.0 1.63 
70–100 16.5 63.7 9.5 26.9 7.32 5.96 55.2 0.83 
Lithic Leptosols (Rankers) Komovi – Rogame 0–10 47.1 26.1 6.8 67.1 3.97 3.39 0.0 16.48 
10–35 24.9 12.2 6.9 71.0 4.29 3.37 0.0 11.32 
Soil group Location Soil depth (cm) > 2 mm 0.02–0.002 < 0.002 2–0.02 (%) pH H2pH KCl CaCO3 (%) Organic matter (%) 
Colluvic Regosols Mrcevopoljea 0–30 10.3 40.7 9.6 49.7 7.09 6.39 4.5 2.99 
30–60 8.5 35.3 11.6 53.3 6.99 6.14 1.9 1.98 
70–100 3.9 42.3 15.4 42.3 7.34 6.54 4.3 1.64 
Rendzic Leptosols Previsa 0–30 19.4 39.1 17.8 43.1 7.7 6.7 13.7 9.1 
Mollic Leptosols Zagricab 0–10 52.6 29.7 6.4 63.9 7.6 6.65 13.7 17.03 
10–20 61 36.0 4.0 60.0 7.59 6.45 23.3 16.08 
Dystric Cambisols Kosuticac 1–14 27.9 43 13.1 43.9 4.97 4.11 0.0 6.08 
14–47 23.9 38.9 18.7 42.4 5.26 4.32 0.0 2.52 
50–70 25.2 39.2 23.1 37.8 5.21 4.36 0.0 2.22 
Eutric Cambisols M. Zdrebaonikd 0–26 43.1 8.4 48.5 6.72 5.26 0.0 3.75 
30–60 41.6 28.1 30.3 6.72 5.71 0.0 1.63 
70–100 16.5 63.7 9.5 26.9 7.32 5.96 55.2 0.83 
Lithic Leptosols (Rankers) Komovi – Rogame 0–10 47.1 26.1 6.8 67.1 3.97 3.39 0.0 16.48 
10–35 24.9 12.2 6.9 71.0 4.29 3.37 0.0 11.32 

aCoastal region; bKarst region; cPolimsko-Ibarski region; dZetsko-Bjelopavlicki region; eMountain region.

Climate change projections

In this study, the impact of climate change on Montenegrin agriculture was assessed by using the projections generated by the regional climate model EBU-POM (Eta Belgrade University Princeton Ocean Model) (Rajković & Đurđević 2010) with a resolution of about 30 km. This model was adopted since it represents a reference for the creation of Montenegrin initial national communication on climate change (Ministry of Spatial Planning and Environment of Montenegro 2010). Three SRES emission scenarios were A1B, A1Bs and A2 and refer to 2001–2030, 2071–2100, and 2071–2100, respectively. According to the concentration of greenhouse gasses, the scenarios are characterized as ‘mean’ (A1B and A1Bs) and ‘high’ (A2) emission scenarios and the expected concentration of CO2 at the end of the 21st century could be about 690 and 850 ppm, respectively.

Modelling tools and simulation methodology

The CROPWAT 8.0 decision support tool (Smith 1992), developed by the Land and Water Division of the FAO, was used to calculate crop water requirements and irrigation needs using climate, soil, crop and management input parameters.

The simulations and data elaborations were conducted for the baseline and three climate projection scenarios. The main methodological steps included: (a) estimate of reference evapotranspiration (ETo) for 35 meteorological stations in Montenegro and eight in the neighbouring countries for the baseline scenario (1961–1990); (b) determination of the length of growing season for the baseline scenario by means of growing degree days (GDD) approach; (c) set up of CROPWAT climate, soil and management files for baseline scenario; (d) estimation of actual evapotranspiration (ETa), NIR, and relative yield (RY) for rainfed cultivation and various irrigation inputs; (e) determination of the length of growing season for the SRES scenarios; (f) set up of CROPWAT climate, soil and management files for the SRES scenarios; (g) estimation of crop evapotranspiration (ETa), NIR, and relative yield (RY) for rainfed cultivation and various irrigation scenarios in the future conditions; (h) comparison and spatial presentation of results and changes between the SRES scenarios and baseline situation.

Reference evapotranspiration (ETo) was estimated by the Penman–Monteith temperature method (PMT) with limited weather parameters (Allen et al. 1998; Todorovic et al. 2013), for both baseline and future climate scenarios. Crop evapotranspiration (ETc) was estimated on a monthly basis as a product of ETo and crop coefficient (Kc) obtained from the FAO 56 (Allen et al. 1998). NIR were determined by CROPWAT applying a simple soil water balance model. Water stress occurred each time the soil moisture content was depleted below the readily available water content, which was set up to 35% of the total available water for all potato growing stages. The reduction of yield due to water stress was estimated by the Stewart's form of water-yield model (Stewart et al. 1977).

The results of CROPWAT simulations obtained at 43 locations were inserted in ArcGIS, elaborated, and presented by thematic maps, which referred to NIR, crop evapotranspiration, and yield reduction.

Temperature-driven growth

The temperature-driven growth was used to determine the duration of phenological stages for potato. GDD were used to measure the heat energy accumulation that a crop can use for growth and development. The growth stage of crop related to crop water response is determined. Each phenological stage (i.e. emergence, flowering, maturity) is followed by a certain sum of temperatures required for crops to enter that stage. The crop growth stages correspond to a single crop coefficient curve used for evapotranspiration estimate. Temperature sums (GDD) are calculated by means of base temperature (Tb) of 6 °C and cutoff temperature (Tcutoff) of 30 °C (Stockle et al. 2003) using the following equation (Ritchie 1991): 
formula
(1)
where GDD = growing degree days (°C-days); n = number of days; Tavg = daily mean air temperature (°C), obtained as the average between daily maximum and daily minimum temperature (if Tavg<Tb, then Tavg=Tb and if Tavg>Tcutoff, then Tavg=Tcutoff); Tb = base temperature (°C); Tcutoff = cutoff temperature (°C).

In this study, the impact of cut-off temperature was neglected because the average maximum monthly temperatures did not approach the cut-off temperature. This is a disadvantage of the methodology based on monthly time step calculation procedures. However, due to the resolution and availability of input weather data, the monthly time step computation procedures were the most adequate solution. Sum of GDD was calculated in excel spreadsheets on a monthly time step. This is a disadvantage of the GDD calculation on a monthly basis because it disables refinement of the simulations over the season. Taking into consideration potato cultivation, this concept was used for prediction of all the growth stages. Nevertheless, in the case of some meteorological stations located in cold climates (i.e. Žabljak) which have a shorter vegetation period, this concept was used jointly with the last autumnal day when harvest is still enabled. Accordingly, a cultivar with minor growing potential or earlier harvesting date was assumed, which required the correction of Kc values for the last vegetation stage in water balance calculation.

The GDD sums were determined for a common potato cultivar (‘Kennebec’) that covers more than 60% of production in Montenegro. The GDD sums were estimated for both baseline conditions and the SRES scenarios. In the case where the crop could not reach the requested GDD sum in the baseline conditions, the other GDD-formula was used in computation. This formula referred to the choice of lower quality or quantity cultivar. These simple adjustments were done for the Žabljak, Njegovuđa, Rožaje and Dragaši areas. The expected result of these changes is prolongation of vegetative growth, which, on one side, could improve the quality and increase the quantity of yield, and, on the other side, requires a higher amount of agricultural input, mainly fertilizers and water. The analyses for the SRES scenarios were carried out using the ‘dumb-farmer approach’ (Rosenberg 1992; Knezevic et al. 2017), which means that the only changing factor in the future production is climate. The other factors are assumed to remain constant.

Spatial data analysis

The results of soil water balance simulations were interpolated in the GIS environment (ESRI 2011) applying the exact interpolator in calculation procedures. The spatial presentation of data was carried out by adopting a completely regularized spline interpolation technique. This method estimates values using a mathematical function that minimizes overall surface curvature, resulting in a smooth surface that passes exactly through the input points (Jeffrey et al. 2001). The spline interpolation technique was applied in numerous studies at different scales and included several climate and hydrological variables (Todorovic et al. 2013; Knezevic et al. 2017).

RESULTS AND DISCUSSION

Air temperature and precipitation in climate change scenarios

The variation of mean annual temperature across the country for the baseline conditions and three climate projections is presented in Figure 2.

Figure 2

Spatial variation of the average yearly air temperature in Montenegro for: (a) baseline (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

Figure 2

Spatial variation of the average yearly air temperature in Montenegro for: (a) baseline (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

The variations are more pronounced between the regions than between the different scenarios (Figure 2). In the case of the A1B scenario (Figure 2(b)), period 2001–2030, the greatest temperature increase occurred during the summer (JJA) season, with values ranging from +1.3 °C in the north to +1 °C in the coastal zone. In the winter season (DJF), the increase of air temperature was about 0.5 and 0.9 °C in the coastal area and the northern mountainous part of the country, respectively. In the spring season (MAM), the change of temperature was more significant compared to the winter season with values ranging from 0.8 °C in the south to 1.1 °C in the north. The autumn season (SON) was characterized by an almost homogeneous temperature change of about 0.7 °C observed over the whole territory.

In the case of the A1Bs scenario (Figure 2(c)), period 2071–2100, spatial patterns of change were similar to those observed for the period 2001–2030, but with a greater magnitude. The areas along the Adriatic Sea demonstrated a lower change in temperature compared to the areas in the northern mountainous region. This is in agreement with the findings of Saadi et al. (2015). The largest change in temperature occurred in the summer season. In the coastal zone, the increase of temperature was about 2.4 °C, and in the mountain region, in the northern part of the country, it was about 3.4 °C. During the winter season, the temperature gradient from the south towards the north of the country was observed, with an increase in temperature ranging from 1.6 °C in the coastal zone to 2.6 °C in the north. In the spring season, the changes of air temperature ranged from 1.6 to 2.6 °C, while in the autumn season they were approximately 1.6 °C in the coastal zone and 2.4 °C in the north, at the border with Serbia.

For the A2 scenario (Figure 2(d)), period 2071–2100, the highest increase of air temperature occurred in the summer season in the mountainous region, with values above 4.8 °C. In the coastal zone, the temperature increased by 3.4 °C during the same season. In the winter season, the temperature changed by 2.6 °C along the Adriatic Coast and approximately 3.4 °C in the northern parts of the country. During the spring season, the temperature increase ranged from 2.8 to 3.6 °C, whereas the change of temperature in the autumn season was uniform and ranged from 2.6 to 3 °C.

Spatial variation of mean annual precipitation for the baseline conditions and three climate change scenarios is given in Figure 3. The results of the model indicated both negative and positive changes in precipitation amounts, depending on the geographic location and the season. Specifically, an increase in precipitation during the summer months was observed in the central parts of Montenegro, while a similar increase occurred in the spring season in the north-west part of the country that borders with Bosnia and Herzegovina. These positive changes were very small and ranged up to 5% above the baseline scenario. In other parts of Montenegro, the results indicated a reduction of precipitation during the winter–spring season. Furthermore, the spring season was characterized by the highest reduction of precipitation, up to 20%, across almost the entire territory of Montenegro. These findings are in agreement with a prior study by Knezevic et al. (2013).

Figure 3

Spatial variation of mean annual precipitation (mm) in Montenegro for: (a) baseline (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

Figure 3

Spatial variation of mean annual precipitation (mm) in Montenegro for: (a) baseline (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

The lowest change of precipitation was observed in the case of A1B scenario (Figure 3(b)), period 2001–2030. A slight reduction of precipitation of about 10% was noted for most locations. The greatest reduction of precipitation was observed in autumn, which is in agreement with the results of Saadi et al. (2015).

For the A1Bs scenario (Figure 3(c)), period 2071–2100, a reduction of precipitation on a yearly basis was observed everywhere and during all seasons, up to 30% in the northern and coastal zones. The spring season was characterized by a uniform precipitation reduction of approximately 10% over the entire country. A similar precipitation reduction was observed during the summer season in the coastal areas, whereas a negative trend ranged from 15 to 20% in the central and northern regions, respectively. In the autumn season, the projection results indicated a significant reduction of precipitation ranging from 30 to 50%.

In the case of the A2 scenario (Figure 3(d)), period 2071–2100, both an increase and decrease of precipitation in respect to the baseline conditions is expected. In the winter season, a positive precipitation trend, ranging from 5 to 10%, was observed only in the north-western parts of the country. In the other parts of Montenegro, the precipitation change during the winter was negative and ranged from 5 to 10%. The most significant precipitation change occurred along the coast in the summer season, with a precipitation decrease of 50%. In this season, the northern parts of the country experienced a precipitation decrease of 10%. In the spring and autumn seasons, the spatial distribution of precipitation was uniform, with a mean decrease of about 20%.

The simulation results at the end of the century (A2 scenario) has foreseen an increase of 4.8 °C in the air temperature during the summer season in the mountain region and 3.4 °C in the coastal zone. In the same scenario, the simulation results indicate a shortening of the vegetation season for almost a month (29.7 days). The strong variation of the agroecological zones and other environmental factors and the extreme rise of the air temperature in the end of the century has already been seen (Knezevic et al. 2017).

Impact of climate change on the length of potato growing season

The simulation results indicated a shortening of potato vegetation season for all SRES scenarios and regions. The results are reported in Figure 4. The growing season started from the end of February in the coastal zones, to the middle of May in the mountainous area of Žabljak. The length of the growing season for the baseline conditions ranged between 116.1 ± 3.0 days in the coastal zone and 124.9 ± 8.6 days in the Karsts region. The average anticipation of vegetation season is predicted as 13.6 days for the A1B scenario, 22.9 days under the A1Bs scenario, and 29.7 days under the A2 scenario. The differences among the regions are high and they are a reflection of the different warming during the spring–summer season across the country. An average decrease of vegetation season for the A1B scenario was around 6 days in the Coastal and Zeta regions, whereas in the Karsts and Mountainous region (Mountainous and Ibarsko-Polimski region) it was 12.6 and 18.8 days, respectively. Scenario A1Bs was warmer and the potato crop is expected to have a shorter growing season of 12.0 days in the Zeta and Coastal regions, 25.8 days in the Karsts region, and 28.1 days in the Mountainous region. The A2 scenario was the warmest, and the potato crop had an additional reduction of the vegetative season of about 6–7 days across the whole country. The greatest difference is foreseen to occur in the mountainous location, where the season starts in May, due to the predicted temperature increase. The lowest difference in the length of growing season was observed in the area of Zeta and in the Coastal region. According to the raster GIS data, the average length of potato vegetation season for the baseline conditions was 122.9 ± 5.0 days, whereas in the future scenarios a decrease is foreseen as 11.5 ± 2.7 days, 24.0 ± 6.2 days and 31.0 ± 6.7 days for A1B, A1Bs and A2, respectively. The overall results are in agreement with the study of Stričević et al. (2017) that showed a similar effect of temperature increase on the shortening of the potato growing season in Bosnia and Herzegovina.

Figure 4

The difference in length of vegetation season for the potato crop grown in Montenegro compared to the baseline scenario: (a) A1B scenario, (b) A1Bs scenario and (c) A2 scenario.

Figure 4

The difference in length of vegetation season for the potato crop grown in Montenegro compared to the baseline scenario: (a) A1B scenario, (b) A1Bs scenario and (c) A2 scenario.

The concept adopted in this study is based on GDD sums, and therefore the prevailing assumption is that the same cultivars require the same GDD in different locations. In practice this means that the cultivars which require a longer growing season could have greater yield and, probably, better quality products. So, in terms of quantity and quality, the GDD sums contribute to specific agricultural systems. It implies that the late maturation cultivars require a longer vegetation period and, therefore, higher water and fertilization requirements, which result in increased quantity of final product and improved quality.

Crop water requirements and irrigation needs

The crop water requirements of potato are presented in Table 2 for the baseline and three future scenarios by means of the absolute values of maximum (ETm) and actual (ETa) crop evapotranspiration corresponding to full irrigation and rainfed cultivation, respectively. The impact of climate change on ETm and ETa is presented in Table 3, by means of percentual change in respect to the baseline conditions. In the case of baseline conditions, ETm over the regions ranged from 359.3 mm in the coastal area to 491.3 mm in the Polimsko-Ibarski region, which represented a variation of almost 37%. Crop evapotranspiration in the coastal area was in the range of values obtained in another study (Cantore et al. 2014) in the coastal area of the nearby Apulia region, southern Italy. Under rainfed cultivation, ETa was in the range of 297.9 mm in the Polimsko-Ibarski region to 325.5 mm in the mountain region. Therefore, the results of ETa were much more homogenous around the country. The greatest difference between ETm and ETa was observed in the Polimsko-Ibarski region which corresponds to the highest yield losses under rainfed cultivation.

Table 2

Potato crop seasonal maximum (ETm) and actual (ETa) evapotranspiration in mm in different Montenegro regions for baseline (1961–1990) and three future scenarios (A1B, 2000–2030; A1Bs, 2071–2100; A2, 2071–2100)

Scenario Baseline
 
A1B
 
A1Bs
 
A2
 
Region ETm ETa ETm ETa ETm ETa ETm ETa 
Mountain 449.1 325.5 399.4 303.4 361.2 234.6 344.0 236.2 
Ibarsko Polimski 491.3 297.9 457.8 276.5 420.6 220.2 398.7 215.8 
Karst 455.4 317.3 422.4 280.5 388.4 240.3 367.8 226.4 
Zeta 380.4 319.4 365.4 292.1 352.9 299.2 334.4 286.4 
Coastal 359.3 302.9 343.9 275.4 329.8 281.7 310.7 266.8 
Scenario Baseline
 
A1B
 
A1Bs
 
A2
 
Region ETm ETa ETm ETa ETm ETa ETm ETa 
Mountain 449.1 325.5 399.4 303.4 361.2 234.6 344.0 236.2 
Ibarsko Polimski 491.3 297.9 457.8 276.5 420.6 220.2 398.7 215.8 
Karst 455.4 317.3 422.4 280.5 388.4 240.3 367.8 226.4 
Zeta 380.4 319.4 365.4 292.1 352.9 299.2 334.4 286.4 
Coastal 359.3 302.9 343.9 275.4 329.8 281.7 310.7 266.8 
Table 3

Differences (% diff.) in maximum (ETm) and actual (ETa) evapotranspiration for winter potato crop in a different regions of Montenegro between three different climate change scenarios (A1B, 2000–2030; A1Bs, 2071–2100; A2, 2071–2100) in respect to baseline scenario (1961–1990)

Scenario A1B
 
A1Bs
 
A2
 
Region ETm ETa ETm ETa ETm ETa 
Mountain −12.4 −6.8 −24.3 −27.9 −30.5 −27.4 
Ibarsko Polimski −7.3 −7.2 −16.8 −26.1 −23.2 −27.5 
Karst −7.8 −11.6 −17.2 −24.3 −23.8 −28.7 
Zeta −4.1 −8.5 −7.8 −6.3 −13.8 −10.3 
Coastal −4.5 −9.1 −8.9 −7.0 −15.6 −11.9 
Scenario A1B
 
A1Bs
 
A2
 
Region ETm ETa ETm ETa ETm ETa 
Mountain −12.4 −6.8 −24.3 −27.9 −30.5 −27.4 
Ibarsko Polimski −7.3 −7.2 −16.8 −26.1 −23.2 −27.5 
Karst −7.8 −11.6 −17.2 −24.3 −23.8 −28.7 
Zeta −4.1 −8.5 −7.8 −6.3 −13.8 −10.3 
Coastal −4.5 −9.1 −8.9 −7.0 −15.6 −11.9 

In the case of the A1B scenario, ETm is foreseen to be reduced from 4 to 12% depending on the region (Table 3). The highest decrease is expected in the Mountainous region. ETa is predicted to decrease in all regions from 6.8 to 11.6% under the A1B scenario.

Much greater reductions are foreseen under the A1Bs scenario. Crop water requirements to reach optimal yield are expected to decrease from around 8% in the Zeta and Coastal regions to 24% in the Mountainous region. Under the A1Bs scenario, ETa is expected to diminish from around 7–8% in the areas characterized by the Mediterranean climate, to 24–28% in the other three regions (Mountain, Ibarsko-Polimski, Karst) that have temperate continental and mountainous climates.

The impact of climate change was the most severe under the A2 scenario. The relative reduction of maximum ETa was 6–7% lower than in the previous scenario, with a 13–15% decrease in the Mediterranean areas, and up to 30% in the mountain zones. The changes in actual water consumption were, although in absolute terms, somehow smaller than in the scenario A1Bs. The actual evapotranspiration decrease, compared to the A1Bs, was the highest in the Mediterranean climate zone. Compared to the baseline scenario, ETa was reduced between 10 and 12% in the coastal areas, whereas the decrease of ETa was about 28% in all other regions.

Spatial variation of NIRs for potato crop are presented in Figure 5, whereas spatial presentation of absolute difference in NIR is given in Figure 6. The impact of climate change on NIR of potato grown in different regions of Montenegro is presented in Table 4. The average NIR to reach maximum yield changed more over the country than between the scenarios. The overall increase of NIR is foreseen as 4.0 mm in the A1B scenario, 19.5 mm in the A1Bs scenario, and 7.3 mm in the A2 scenario. Therefore, the increase was the highest for the A1Bs scenario (14%). The difference among the regions and locations was high. The average lowest NIR were 94.7 and 98.8 mm in the Coastal and Zeta regions, 117.0 mm in the Mountain region, 165.6 mm in the Karst region, and the maximum NIR (215.8 mm) were in the Polimsko-Ibarski region. Compared to the baseline conditions, the Mountain region is expected to have a reduction of NIR of 3% for the A1B scenario, whereas an increase of NIR of 29% is foreseen under the scenario A1Bs. This is partially due to the possibility of growing the longer season cultivars in the locations of Žabljak and Njegovuđa. An average increase of NIR of about 13.0% was also obtained in the A2 scenario. The Polimsko-Ibarski region had the highest NIR but in the future this increase is predicted to fluctuate from a 6% increase under the A1Bs scenario to a 2–5% decrease under the other two scenarios. The Karst region is expected to face an increase in NIR of 8–19% in the future, being the highest for the A1Bs scenario.

Figure 5

Spatial variation of net irrigation requirements (NIR) for potato crop in Montenegro for: (a) baseline conditions (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

Figure 5

Spatial variation of net irrigation requirements (NIR) for potato crop in Montenegro for: (a) baseline conditions (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

Figure 6

Spatial presentation of absolute difference in net irrigation requirements (mm) over the whole Montenegro area for the potato crop compared to baseline scenario: (a) A1B scenario, (b) A1Bs scenario and (c) A2 scenario.

Figure 6

Spatial presentation of absolute difference in net irrigation requirements (mm) over the whole Montenegro area for the potato crop compared to baseline scenario: (a) A1B scenario, (b) A1Bs scenario and (c) A2 scenario.

Table 4

Differences (% diff.) in yield reduction and NIR for winter potato crop in a different Montenegro regions under rainfed conditions between three different climate change scenarios (A1B, 2000–2030; A1Bs, 2071–2100; A2, 2071–2100) in respect to baseline scenario (1961–1990)

Scenario A1B
 
A1Bs
 
A2
 
Region yield NIR yield NIR yield NIR 
Mountain 1.4 −3.2 −10.4 29.2 −6.0 13.0 
Ibarsko Polimski −0.2 −4.8 −8.9 6.6 −7.0 −1.7 
Karst −3.5 8.2 −8.5 19.0 −8.8 16.2 
Zeta −4.6 15.4 1.3 −5.4 1.8 −14.4 
Coastal −4.5 13.4 1.0 0.3 1.9 −4.6 
Scenario A1B
 
A1Bs
 
A2
 
Region yield NIR yield NIR yield NIR 
Mountain 1.4 −3.2 −10.4 29.2 −6.0 13.0 
Ibarsko Polimski −0.2 −4.8 −8.9 6.6 −7.0 −1.7 
Karst −3.5 8.2 −8.5 19.0 −8.8 16.2 
Zeta −4.6 15.4 1.3 −5.4 1.8 −14.4 
Coastal −4.5 13.4 1.0 0.3 1.9 −4.6 

The reduction of seasonal crop evapotranspiration, observed in all three SRES projections, is triggered by the decrease of the length of vegetation season due to global warming and temperature increases. In the areas where rainfed production is prevailing, the effect of climate change on potato cultivation is related to projected rainfall pattern, temperature increase, crop phenology, and soil water holding capacity and depth. In fact, soil characteristics play an important role in water balance simulations and the estimation of irrigation requirements. The area of Montenegro facing the highest precipitation (Karst region) is evidently very scarce in terms of potato cultivation. This is due to shallow soil depth and low water holding capacity of Mollic Leptosols and Rendzic Leptosols which characterize the Karst region. Thus, deep percolation is very large, and only a small portion of total precipitation is effectively used by the crop. The same refers to Lithic Leptosols located in the Mountain region, where the amount of annual precipitation is much lower. The most fertile soils in Montenegro are actually the deepest soils, such as Eutric Cambisols and Colluvic Regosols, or on higher terrains Dystric Cambisols. These soils have higher water holding capacities, and better possibilities for potato production.

Impact of climate change on potato yield

The reduction of potato yield under rainfed cultivation in respect to full irrigation and optimal production is presented in Figure 7. The reduction of potato yield under the baseline conditions was around 17–18% in the Coastal and Zeta region, 33.1 and 27.5% in the Karst and Mountain region, respectively, and 43.6% in the Polimsko-Ibarski region. The overall reduction of potato yield over the territory of Montenegro under baseline conditions was around 29.6 ± 8.1%.

Figure 7

Potato yield reduction in Montenegro for: (a) baseline conditions (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

Figure 7

Potato yield reduction in Montenegro for: (a) baseline conditions (1961–1990), (b) A1B (2001–2030), (c) A1Bs (2071–2100) and (d) A2 (2071–2100) scenarios.

In the A1B scenario, the yield reduction was slightly greater, about 31.0 ± 8.2%. The greatest yield reduction is foreseen for the A1Bs scenario, 36.3 ± 11.6%, thus with the highest deviation. In the case of the A2 scenario, the reduction of potato yield was 34.1 ± 10.9%. Nevertheless, these changes were spatially very large and differed among the regions and climate change scenarios.

Under the A1B scenario, the yield loss was around 4% higher than under baseline conditions in the Coastal, Karst and Zeta region, while there were no relevant changes in the Mountain and Polimsko-Ibarski regions. Under the scenarios A1Bs and A2, an additional yield decrease is foreseen of around 8–10% and 6–9%, respectively, in the Mountain, Polimsko-Ibarski and Karst regions, without any relevant distinction among them in terms of highest/lowest reduction. In the case of A1Bs and A2, it is expected that the regions with Mediterranean type of climate will not face any yield reduction compared to the baseline conditions. On the contrary, a small increase in yield of around 1% is predicted.

The impact of climate change on yield reduction is presented in Table 4. The values refer to the percentage difference between yield under baseline conditions and three SRES scenarios. Under the A1B scenario, the yield reductions of potato were obtained in all zones, except the Mountain region. These reductions are the highest in the coastal area and in the Zeta region. In these regions, yield reduction was slightly smaller under the other two scenarios. Under the A1Bs scenario, the additional yield reductions were the highest in the Karst, Polimsko-Ibarski and Mountain regions, ranging from 8.5 to 10.4%. Under the A2 scenario, the reduction of yield in the Mountain region was 6.0%, and it was the highest in the Karst region (8.8%).

The results pointed out that the impact of climate change could lower the yield of potato in the coastal regions, whereas it could be enhanced in the mountain areas. Therefore, a gradual moving of production to the northern and colder areas of the country is expected. In the future, potato could also be successfully grown in the Zeta and Coastal regions, whereas potato growers in the interior part of the country should consider the adaptation measures such as earlier sowing in order to avoid heat stress and lack of available soil moisture in the summer period. A large fluctuation in the yield could be observed and a large decline in the production could be expected. Similar findings on potato productivity have been reported in the literature (Daccache et al. 2011; Cantore et al. 2014; Raymundo et al. in press). The adoption of strategies could attenuate the negative impacts of climate change on potato production (Raymundo et al. 2017).

CONCLUSIONS

The climate change impact on the cultivation of potato was investigated with a simplified GIS based methodology and soil water balance model adapted to the national scale. The results obtained for the baseline condition have been shown to fit quite well with the literature data. Certainly, a more accurate assessment could be possible using local, site-specific crop phenological parameters and water balance models. This is especially important due to the great heterogeneity of soil characteristics in Montenegro.

The combined effects of temperature increase and precipitation decrease would lead to the depletion of the available soil water in the root zone and water stress under rainfed cultivation in the future. Rainfed cultivation remains one of the viable solutions, especially when the anticipation of sowing is adopted as an adaptation measure. In the future, potato cultivation in Montenegro should be shifted to higher altitudes and in less favourable conditions. This means the development of new agricultural areas and adoption of a series of adaptation measures that includes change of cultivars, improvement of crop and soil management and altered crop rotation with the scope to improve water use efficiency and enhance potato production in the future. Otherwise, climate change could have a negative impact on potato production in this region.

Further research efforts should be devoted to estimating water use and productivity in a more sophisticated manner using denser rasterized soil data, DEM's and dual Kc approach. The accuracy of simulation could be enhanced with: (a) detailed meteorological data sets enabling the use of the standard Penman–Monteith equation, (b) availability of daily weather data and the latest climate change projections, (c) use of more detailed and updated soil maps, (d) application of crop growth models including the impact of CO2 on biomass and yield and presenting results in absolute terms, and (e) considering the future land use evolution and a dynamic link between data and models in GIS.

REFERENCES

REFERENCES
Allen
R. G.
,
Pereira
L. S.
,
Raes
D.
&
Smith
M
, .
1998
Crop Evapotranspiration. Guidelines for Computing Crop Water Requirements
.
FAO Irrigation and Drainage Paper 56
,
FAO
,
Rome
.
Burić
D.
,
Ducić
V.
&
Mihajlović
J
, .
2014
The climate of Montenegro: modificators and types, part two
.
Bull. Serbian Geogr. Soc.
9
,
73
90
.
Callaway
J. M.
,
Kašćelan
S.
&
Marković
M
, .
2010
The Economic Impacts of Climate Change in Montenegro: A First Look
.
UNDP
,
Podgorica, Montenegro
.
Cantore
V.
,
Wassar
F.
,
Yamac
S. S.
,
Sellami
M. H.
,
Albrizio
R.
,
Stellacci
A. M.
&
Todorovic
M
, .
2014
Yield and water use efficiency of early potato grown under different water regimes
.
Int. J. Plant Prod.
8
(
3
),
409
428
.
Čereković
N.
,
Todorović
M.
,
Snyder
R. L.
,
Boari
F.
,
Pace
B.
&
Cantore
V.
2010
Evaluation of the crop coefficients for tomato crop grown in a Mediterranean climate
.
Options Méditerranéennes A
95
,
91
94
.
Daccache
A.
,
Weatherhead
E.
,
Stalham
M.
&
Knox
J
, .
2011
Impacts of climate change on irrigated potato production in a humid climate
.
Agr. Forest Meteorol.
151
(
12
),
1641
1653
.
ESRI
2011
ArcGIS Desktop: Release 10
.
Environmental Systems Research Institute
,
Redlands, CA
.
Fuštić
B.
&
Đuretić
G.
2000
Soils of Montenegro
.
University of Montenegro
,
Podgorica
.
Giannakopoulos
C.
,
Le Sager
P.
,
Bindi
M.
,
Moriondo
M.
,
Kostopoulou
E.
&
Goodess
C. M
, .
2009
Climatic changes and associated impacts in the Mediterranean resulting from a 2°C global warming
.
Glob. Plan. Change
68
(
3
),
209
224
.
Jeffrey
S. J.
,
Carter
J. O.
,
Moodie
K. B.
&
Beswick
A. R
, .
2001
Using spatial interpolation to construct a comprehensive archive of Australian climate data
.
Environ. Model. Softw
.
16
,
309
330
.
Jovovic
Z.
,
Micev
B.
&
Velimirovic
A
, .
2016
Impact of climate change on potato production in Montenegro and options to mitigate the adverse effects
.
Acad. J. Environ. Sci
.
4
(
3
),
47
54
.
Knezevic
M.
,
Perovic
N.
,
Zivotic
Lj.
,
Ivanov
M.
&
Topalovic
A
, .
2013
Simulation of winter wheat water balance with Cropwat and Isareg models
.
Agric. Forest.
59
(
1
),
41
53
.
Knezevic
M.
,
Zivotic
Lj
,
Perovic
V.
,
Topalovic
A.
&
Todorovic
M
, .
2017
Impact of climate change on olive growth suitability, water requirements and yield in Montenegro
.
Ital. J. Agrometeorol.
2
,
39
52
.
Lovelli
S.
,
Perniola
M.
,
Di Tommaso
T.
,
Ventrella
D.
,
Moriondo
M.
&
Amato
M
, .
2010
Effects of rising atmospheric CO2 on crop evapotranspiration in a Mediterranean area
.
Agric. Water Manage
.
97
(
9
),
1287
1292
.
Ministry of Spatial Planning and Environment of Montenegro
2010
The Initial National Communication on Climate Change of Montenegro to the United Nations Framework Convention on Climate Change (UNFCCC)
.
Podgorica
,
Montenegro
.
Rajković
B.
&
Đurđević
V.
2010
The report on nesting EBU-POM model into global simulation SX-G model. The Initial National Communication on Climate Change of Montenegro to the United Nations Framework Convention on Climate Change. Ministry for Spatial Planning and Environment of Montenegro, Montenegro
.
Raymundo
R.
,
Asseng
A.
,
Prasad
R.
,
Kleinwechter
U.
,
Concha
J.
,
Condori
B.
,
Bowen
W.
,
Wolf
J.
,
Olesen
J. E.
,
Dong
Q.
,
Zotarelli
L.
,
Gastelo
M.
,
Alva
A.
,
Travasso
M.
,
Quiroz
R.
,
Arora
V.
,
Graham
W.
&
Porter
C
, .
2017
Performance of the SUBSTORpotato model across contrasting growing conditions
.
Field Crops Res
.
202
,
57
76
.
Raymundo
R.
,
Assenga
S.
,
Robertson
R.
,
Petsakos
A.
,
Hoogenbooma
G.
,
Quiroz
R.
,
Hareau
G.
&
Wolf
J.
,
In Press
Climate change impact on global potato production
.
Eur. J. Agron.
https://doi.org/10.1016/j.eja.2017.11.008.
Ritchie
J. T.
1991
Wheat phasic development
. In:
Modeling Plant and Soil Systems
(
Hanks
J.
&
Ritchie
J. T.
, eds).
American Society of Agronomy Inc., Crop Science Society of America, Inc., Soil Science Society of America, Inc
.
Madison, Wisconsin
,
USA
, pp.
31
54
.
Rodriguez-Diaz
J. A.
,
Weatherhead
E. K.
,
Knox
J. W.
&
Camacho
E
, .
2007
Climate change impacts on irrigation water requirements in the Guadalquivir river basin in Spain
.
Reg. Environ. Change
7
(
3
),
149
159
.
Rosenberg
N. J
, .
1992
Adaptation of agriculture to climate change
.
Clim. Change
21
,
385
405
.
Saadi
S.
,
Todorovic
M.
,
Tanasijevic
L.
,
Pereira
L. S.
,
Pizzigalli
C.
&
Lionello
P.
2015
Climate change and Mediterranean agriculture: impacts on winter wheat and tomato crop evapotranspiration, irrigation requirements and yield
.
Agric. Water Manage.
147
,
103
115
.
Smith
M
, .
1992
CROPWAT, A Computer Program for Irrigation Planning and Management
.
FAO Irrigation and Drainage Paper 46
.
FAO
,
Rome
.
Stewart
J. I.
,
Hagan
R. M.
,
Pruitt
W. O.
,
Danielson
R. E.
,
Franklin
W. T.
,
Hanks
R. J.
,
Riley
J. P.
&
Jackson
E. B.
1977
Optimizing Crop Production Through Control of Water and Salinity Levels in the Soil
.
Reports Paper 67
.
Utah Water Research Laboratory
,
USA
, p.
191
.
Stockle
C. O.
,
Donatelli
M.
&
Nelson
R
, .
2003
Cropsyst: a cropping systems simulation model
.
Eur. J. Agron
.
18
,
289
307
.
Stöckle
C. O.
,
Nelson
R. L.
,
Higgins
S.
,
Brunner
J.
,
Grove
G.
,
Boydston
R.
,
Whiting
M.
&
Kruger
C.
2010
Assessment of climate change impact on Eastern Washington agriculture
.
Climatic Change
102
(
1–2
),
77
102
.
Stričević
R.
,
Trbić
G.
,
Vujadinović
M.
,
Cupać
R.
,
Đurović
N.
&
Ćosić
M.
2017
Impact of climate change on potato yield grown in different climatic zone in Bosnia and Herzegovina
.
Conference: VIII International Scientific Agriculture Symposium ‘AGROSYM 2017’
,
Jahorina, Bosnia and Herzegovina
, pp.
596
601
.
Supit
I.
,
van Diepen
C.
,
de Wit
A.
,
Wolf
J.
,
Kabat
P.
,
Baruth
B.
&
Ludwig
F
, .
2012
Assessing climate change effects on European crop yields using the Crop Growth Monitoring System and a weather generator
.
Agric. Forest Meteorol.
164
,
96
111
.