Suitable areas for cultivation of protein-rich crops in Sweden : An GIS-based study on 7 protein-rich crops

(1)

Department of Thematic Studies Environmental Change

MSc Thesis (30 ECTS credits) Science for Sustainable development

Fanny Cardegård

Suitable areas for cultivation of

protein-rich crops in Sweden

An GIS-based study on 7 protein-rich crops

(2)

ii

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility.

According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.

For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page:

http://www.ep.liu.se/. © Fanny Cardegård

(3)

(4)

iv

Abstract

Conditions are predicted to become more favorable for protein-rich crops in Northern Europe, which bring opportunities for cultivation of protein-rich crops in Sweden. This study assessed the current suitability for cultivation of protein-rich crops in Sweden based on biophysical conditions. This study defined suitability indicators with suitability conditions for each crop regarding (i) soil texture, (ii) soil pH value, (iii) vegetation period length, (iv) and soil temperatures were created, and overlayed to create suitable areas for protein-rich crops. Suitable areas and distributions were found for: common bean, faba bean, field pea, lentil, narrow-leafed lupin, quinoa, and soybean. A present and future risk analysis with wet and dry periods was carried out to find suitable areas for the crops under the risk of wet and dry periods. The study found that Sweden have the possibility to cultivate protein-rich crops to a greater extent which is shown by the distribution of suitable areas for protein-rich crop. Quinoa was found suitable to be cultivated in nearly all arable land in Sweden. In the future, there is an increase in suitable areas for protein-rich crops that are not exposed to drought. A decrease was seen in suitable areas for protein-rich crops that are not exposed to flooding.

Key words: drought, flooding, legumes, protein-rich crops, pseudo-cereals.

1 Introduction

In order to sustainably supply food to a growing population it is not enough to only increase productivity, improve management and new technologies. A transition is needed towards diets that demand less resources and contain less animal products (Röös et al., 2018). A way to do this is to eat more plant-based food in the direction of reducing environmental pressures from the food system (Röös et al., 2018). To replace animal protein with plant protein diets have been proven to give both health and environmental benefits, leading to interest in the potential of protein-rich crops as an alternative for animal products. Examples of such protein-rich crops are legumes and pseudo-cereals. For more than two decades, the production and consumption of protein-rich crops have declined in many parts of the world, including Europe, the cause is both agronomic and socio-economic (Manners et al., 2020).

Globally, grain legumes are mostly used as feed for livestock, but grain legumes are also used for direct human consumption. Grain legumes exhibit a critical function important from an agronomic perspective because of their ability to fix nitrogen in symbiosis with soil living bacteria. As pre-crops, grain legumes play an essential part in bringing nitrogen in the crop rotation, by leaving nitrogen in the soil for the following crop (Döring., 2015). They also bring other effects for the following crop, such as helping to break the life cycle of soil-borne diseases of cereals, for example the disease take-all, caused by the pathogen Gaeumannomyces graminis var. tritici, which wheat is sensitive to (Döring., 2015; Kirkegaard et al., 2008). One other important function of grain legumes are the provision of nectar and pollen resources for pollinators (Döring., 2015). In some European countries, recent reversals in legume production is occurring. There are signs of an upcoming increase in protein-rich crop production is due to new growing demand, increasing consumer awareness of the health benefits of dietary transitions, and long-term high prices of plant proteins that is expected (Manners et al., 2020). With increasing production of protein-rich crops it will be necessary to address the challenges that comes with climate change. The estimated increase in global average temperature of 1.5 °C by 2100 will lead to great impacts on global crop productivity which in turn affects crop distribution. Models predicts that in the future, conditions in northern Europe will become more favorable for protein-rich crops. Climate change could therefore lead to an increased production

(6)

2

potential of protein-rich crops in Northern Europe, which in addition to a shift from animal to plant-based protein, also may support food supply at a regional level (Manners et al., 2020). Sweden might gain importance as an agricultural producer in the future and be able to contribute more to the global production, as agricultural production in many other countries could decrease, since they are negatively affected by climate change (Juhola et al., 2017).

Climate change is predicted to bring wetter and warmer conditions for Nordic countries, leading to for example increased risks of infestation of pest and weed. Periods of drought are expected to be longer and days with heavy precipitations are predicted to increase, which are likely to have apparent effect on agriculture. Altered conditions for crop production as result of climate change may thus pose both challenges and opportunities for future agriculture (Neset et al., 2019). Drought and heavy precipitation that could lead to flooding are two abiotic stresses that affect crop growth. However, different crops respond and adapt differently to stresses (González et al., 2009), which signifies different vulnerabilities to climate change and different suitability for crops in a changed climate.

1.1 Problem formulation

There is a need for increasing production of protein-rich crops as an alternative for animal products due to a growing population worldwide, and increased health and environmental benefits (Röös et al., 2018). The production of legumes in Europe have declined as the importation of soybean has increased (Bodner et al., 2018). There are several agronomic advantages with legumes such as fixing nitrogen and working as a break crop to soil-borne diseases (Döring., 2015). It is therefore of interest to identify the current potential of growing more protein-rich crops in Sweden. With a changing climate it is likely that the suitability for protein-rich crops will improve, however, since the crops are sensitive to both drought and flooding it is valuable to assess how the crops can be exposed to these stressors, today and under climate change.

The aim of this thesis is to assess the current suitability for cultivation of protein-rich crops in Sweden based on biophysical conditions. The risk of wet and dry periods over current and future periods will be assessed in relation to the suitable areas. A number of different protein-rich crops will be analysed: common bean, faba bean, field pea, lentil, narrow-leafed lupin, soybean, and quinoa. The study aims to identify suitable areas for these crops and assess the potential distributions of them in Sweden. To do that, the study examines some conditions that needs to be fulfilled to produce protein-rich crops suitable for Sweden. A present and future analysis based on climate simulations with wet and dry periods will be carried out to find suitable areas for the crops under the risk of wet and dry periods. The study will be focused on southern to the middle of Sweden excluding the northern parts, since there are poorer farming conditions and less data up north.

The research questions are as followed:

• Where and how large are the current suitable areas for cultivation of the chosen protein-rich crops in Sweden?

• Where in Sweden under current and future climates are the identified suitable areas for protein-rich crops not exposed to wet and dry periods?

(7)

3

2 Background

2.1 The potential of growing protein-rich crops

Possible alternative products to animal products are protein-rich crops, such as legumes and pseudo-cereals (Manners et al., 2020). Legumes have its place in the plant family called Fabaceae. In this plant family, crops are for example grown for their seeds (dry beans, dry peas and lentils). The family also includes fresh vegetables such as green beans and green peas. Moreover, the family contain livestock forage such as clover and alfalfa. Legumes offer valuable agronomic advantages, where one of them is symbiotic nitrogen (N) fixation and another is the use of legumes as break crops in cereal based cropping systems (Röös et al., 2018).

The large-seeded legumes are called grain legumes or pulses and they are cultivated for the harvesting of mature seeds (Fogelfors, 2015, p. 299). In the near future, the demand for grain legumes in the world is expected to increase. This increase is expected not only in developing countries but in de developed countries as well due to an increasing trend towards healthy dieting. The demand for legume-based products will keep increasing since legumes have been found to have many therapeutic uses and the health risk of animal protein consumption is broadly acknowledged. Most legumes are rich in proteins and soluble fiber. Regular intake of legumes is correlated to reduction in risk of cardiovascular diseases, diabetes, digestive tract diseases, and obesity (Daryanto et al., 2015). Grain legumes are of vital significance as a source of protein. Grain legumes can be seen as a vital complementary function for nutrition for both human and animals, especially in contrast to cereals. However, a stable decline in area grown with grain legumes over the last years have been seen for many countries. This is due to a dramatic change in subsidies for growing grain legumes in the European Union. Other factors such as the use of mineral N fertilizer and replacement of grain legumes in human consumption to meat consumption are also affecting the decline. At a global scale, there is significant expansion of areas over the last two decades which is due to cultivation of soybean which has increased by 64.4 % between 1993 and 2012 (Döring., 2015). The legume production area in Europe has decreased around 67 % between the years 1961 to 2014. The European decrease in legume productions is related to a more than a 6-fold increase in soybean imports during the same time (Bodner et al., 2018). The extensive import of soybean as a replacement for domestic production of protein crops legumes in particularly, is not sustainable (Zander et al., 2016). The Food and Agriculture Organization of the United Nations (FAO) promoted the international Year of Pulses in 2016 in order to use pulses to eradicate hunger and malnutrition in the world. The initiative focused on promoting the value and usage of pulses in the food system, enlightenment of benefits, encourage enhanced research, promote better use of pulses in crop rotation, and tackle trade challenges. The initiative was also linked to the Sustainable Development Goal 2 and how pulses can contribute to food access, malnutrition, and to agriculture that are sustainable and resilient (Considine et al., 2017).

An increased global production of grain legumes has the potential to offer a sustainable solution to secure food and protein. Now ongoing work is being made to increase genomic resources and improvement of yield and nutritional quality of legumes with innovative breeding techniques, and better resilience to climate change (Considine et al., 2017). In southern Europe, where traditional protein-rich crops today are grown will be less favorable of future climate and it could limit the production. Quinoa will have higher suitability in Europe in the future and be able to expand in area. In Europe, legumes such as faba bean, lentil and chickpea will have to be improved in tolerance for abiotic stresses since these are likely to be more common with climate change (Manners et al., 2020).

(8)

4

2.2 Swedish agriculture

Farmland is an overall concept for arable land, pastures, and mown meadows. The arable land is defined as soil suitable for plowing, or to be used as orchard, nursery-garden, or energy forest. Pastures are the soils that are not suitable for plowing and are therefore used by grazing animals. Mown meadows are farmland that does not have arable soils and are used in late summer for mowing, grazing, or gathering of leaf fodder. In Swedish statistics areas of mown meadows are few and is therefore included in pastures. The farmland in Sweden (2015) were 3 039 896 ha, the arable land consisted of 2 590 052 ha. The farmland area is 7.5 % of the total land area in Sweden (SCB., 2019).

The conditions for agriculture are very different in the north compared to the south, the growing season is for example almost 100 days longer in most southern location compared to the most northern locations. The crop production in Sweden is dominated by cereals (wheat, barley, and oats) and grasslands. Cereals are cultivated on around 40 % of the arable land in Sweden. Crop distribution varies from north to south which is explained by different climate conditions. Forage and coarse grain are cultivated in the north while bread grain is produced in the plains of south and central Sweden. Oilseed production consists mainly of rapeseed and colza located in south and central Sweden. Potatoes is produced in all of Sweden, but sugar beets are only produced in the most southern parts. Carrots and iceberg lettuce are the most important vegetables (Swedish Board of Agriculture., 2009).

2.3 Protein-rich crops in Sweden

Sweden have a long tradition of bean cultivation. Formally it can be followed back to the late 1800s, it is considered that bean cultivation in different forms can be dated back to the 1600s (Fogelberg., 2008). Cultivation of lentils can be traced back to the 1800s on Gotland were the landrace Gotland lentil was cultivated (Nordisk råvara., 2020b). The interest of legumes has increased lately and in Sweden it is mainly peas, faba beans, and, in smaller extent common bean (brown beans), vetches, lupin, soybean and lentils. As greenfodder, often together with cereal is often peas, faba bean, vetches and in smaller extent lupin cultivated (Fogelfors, 2015, p. 299). In 2017 grain legumes were grown on 2.2 % of Sweden’s total cropland area (Röös et al., 2018).

Currently in Sweden, the cultivation of protein-rich crops is produced for food and feed consumption (tab. 1). The crops with the largest production for Sweden are faba bean, pea, and common bean. (Stoddard., 2017). Over the last decade smaller production for food consumption of common bean (brown beans, black beans, kidney beans, white beans, and borlotti, and more) have been produced on the island of Öland (Go green., 2020). There is a current small-scale production of green lentil, red lentil, Gotland lentil, quinoa, and lupin bean cultivated for food consumption in Sweden (Nordisk Råvara., 2020a). Protein-rich crops are also cultivated for fodder production, and most of the produced faba beans is used as animal feed (Röös et al., 2018). Narrow-leafed lupin is cultivated for fodder production (scandinavianseeds., 2020) and is considered to be an alternative crop for soybean (Linnskog Rudh., 2018). Soybean is not being cultivated for production for consumption of food or fodder, because Sweden is currently unable to refine soybean into protein meal (Linnskog Rudh., 2018).

(9)

5

Tab. 1. Cultivation and production of protein-rich crops in Sweden.

Cultivation and production of protein-rich crops in Sweden

Food consumption Cultivated for fodder

Common bean Small-scale production

(Röös et al., 2018).

No (Stoddard., 2017).

Faba bean Production (Röös et al.,

2018).

Yes (Röös et al., 2018).

Field pea Production (Stoddard.,

2017).

Yes (Olrog., 2004).

Lentil Small-scale production

(Nordisk Råvara., 2020a).

No (Stoddard., 2017).

Narrow-leafed lupin Small-scale production

(Nordisk Råvara., 2020a).

Yes (Scandinavianseed., 2020).

Quinoa Small-scale production

(Nordisk Råvara., 2020a). No

Soybean No production (Linnskog

Rudh., 2018).

Not for production (Linnskog Rudh., 2018).

2.3.1 Common bean

Common bean (Phaseolus vulgaris L.) was domesticated in the Mesoamerican and Andean areas around 4000 bc and the domestication occurred independently which gave rise to two different gene pools. Common bean came to Europe at the end of the 15th century and the spreading of common bean was complex. Most European landraces of common bean originates from the Andean gene pool (Angioi et al., 2010; Stoddard., 2017). Common bean is predominately self-pollinating and is the world’s third most important food legume crop after soybean and peanut. It has a great variability in growth habit, seed attributes, and maturation time between different varieties. This variability is what makes the common bean suitable for cultivation in many variated cropping systems and ecosystems around the world (De ron et al., 2015). Beans belong to one of the most nutritious food products with protein content of approximately 20 %, but it varies between varieties and between years. All varieties of bean cultivation require lime conditions generating a high soil pH value. Beans are sensitive of frost during germination and emergence and requires due to that warm soils (Fogelberg., 2008). Common bean is sensitive to water scarcity, waterlogging and salinity (Stoddard., 2017). The practical restraints for domestic bean cultivation in Sweden are considered minor, but the production of common bean is however small. Southeastern parts of Skåne, and the islands of Öland and Gotland are considered suitable for bean production due to early springs and relatively warm autumns (Fogelberg., 2008). The technique for cultivation of beans is well-known and there is a wide selection of international beans suitable for Sweden. One of the Swedish landraces are called brown beans and it is cultivated on Öland since a hundred years back (Fogelberg., 2008). Brown beans from Öland is one of few protected Swedish food

(10)

6

products, and it is protected according to the EU-regulation within the protected geographical indication (PGI) (Olsson., 2015). Common bean is used as a food crop and rarely used as a feed crop (Stoddard., 2017).

2.3.2 Faba bean

The faba bean (Vicia faba L.) is a rich protein grain legume with content reaching from 17.6 -34.5 %. It is belonging to the Fabaceae family and is also referred to as horse bean or broad bean. In northern Europe, tradition of cultivation of faba bean is long, were not much is known about the exact origin or the steps of the domestication process (Duc et al., 2015). In a global perspective faba bean is ranked seventh position among the grain legumes and is placed behind both soya bean and common bean. In a European perspective faba bean is placed second after pea and before soybean. In Europe, faba bean covers 21.3 % of the 1.6 million ha used for grain legume cultivation (Duc et al., 2015).

Faba bean has increased in cultivated areas, which is due to an increased domestic use by the fodder industry (Fogelfors, 2015, p. 216-217).

Faba bean has been widespread since historic times which has resulted in regional adaptations to abiotic stresses and variation in resistance to diseases. In Ecuador, resistance has been found to the disease chocolate spot. In Bangladesh tolerance to heat stress was found. Frost tolerance in winter types of faba bean from Europe were found, and drought tolerance were found in the Mediterranean region (Duc et al., 2015). For cultivation of faba bean some technical problem arises. The equipment used for production can need adjustment or replacement since the equipment used for small grains such as cereal is not suitable for faba bean. This mean that specific investments need to be done for cultivation of faba bean (Duc et al., 2015).

2.3.3 Field pea

Field pea (Pisum sativum) is grown in most of the worlds temperate zones and it belongs to the Leguminosae family and can therefore fix atmospheric nitrogen. Pea is one of the oldest domesticated crops in the world, and in Europe, it has been cultivated since the Stone and Bronze Ages. Field pea is also called dry pea and combining pea (Warkentin et al., 2015). Protein content in field pea is around 23 % (Holstmark., 2016). Field peas have varieties used as food and other used as fodder (Olrog., 2004).

In Europe, field pea breeding increased during the 1980s and the 1990s due to provided subsidies from EU policy for the local production of protein-rich crops. The reason for these subsidies was to address the substantial insufficiency in protein in Europe. It is primarily the private sector that is conducted to pea breeding in Europe while the public sector focuses on upstream research. However, Eu policy changed and became less favorably to protein crops which led to a decline in the production and breeding of field pea ever since the beginning of 2000 (Warkentin et al., 2015).

Many biotic and abiotic stresses must be tackled through breeding before cultivation of peas could achieve good yield. The biotic stresses are fungal diseases, insects, and viruses. Heat stress at flowering and early season flooding are key abiotic stresses for field pea. These stressors are in focus for pea breeders to increase and stabilize yields (Warkentin et al., 2015). Field peas have high demands on the root environment. Humid, sealed, poorly ditched, packed soils often lead to problem with pea root rot, which lead to bad development. Years with high precipitation increases the risk for bad root environment especially on clay soils with bad permeability (Olrog., 2004).

(11)

7

Traditionally, field peas are cultivated more on the eastern side than the western of Sweden, which is corelated to different circumstances regarding soil and precipitation. Field peas should normally not return more often than every 6-8 year to increase the risk of root rot and other diseases. An alternative can be co-cultivation together with oat or barley (Olrog., 2004). 2.3.4 Lentil

Lentil (Lens culinaris Medik.) belongs to the legumes or the Fabaceae family and it was one of the first domesticated grain crops. Protein content in lentil varies from 21.4 to 25.5 % (Stefaniak and McPhee., 2015). It is possible to grow lentils in boreal climates. Lentil gives lower protein yield than other legumes. It is an essential specialty food crop and is suitable for short growing seasons (Lizarazo et al., 2014). Loam and drained lighter soil with dry autumns are suitable for cultivation of lentils. Lentils have a shallow root system and are therefore very sensitive to standing water (Adolfsson., 2013; Stoddard., 2017).

Lentil have been cultivated in Europe since the Bronze Age and it is still important in the Mediterranean region, Western Asia and North America (Stoddard., 2017). Lentil is a food crop, with small-scale production in Sweden (Nordisk Råvara., 2020). Lentil is not used as feed since yields are considered too low and production cost too high. Conventional farm machinery made for handling cereals such as wheat and barley work properly for seeds from lentils which make it easier for cultivation since no adjustments or replacements of machinery is needed (Stoddard., 2017).

2.3.5 Narrow-leafed lupin

Narrow-leafed lupin (L. angustifolius) is a lupin that is mostly grown in northern Europe (Gresta et al., 2017). Narrow-leafed lupin belongs to the species of lupin that origins from the Old World and it belongs to smooth-seeded lupins (Święcicki et al., 2015).

Wild lupins have a chemical defense where they produce quinolizidine alkaloid which are bitter and toxic (Gresta et al., 2017). Decreasing alkaloid content in seeds were the most important step during the domestication of lupins to increase it values for feed and fodder. Seeds of wild lupin have several percent in alkaloid content. There exists no exact level of the alkaloid content that can be used for feed. Currently there are different levels acceptable in different countries. The newest cultivars contain less than 0.01 %. The lupins are divided into two groups after the alkaloid content using the names bitter for high alkaloid content, and sweet for low alkaloid content. The first “sweet” plants of narrow-leafed lupin were emerging during the years 1927-1928 and made it suitable for fodder. It expanded in popularity as a fodder crop since there was as increasing demand for protein feed for pigs and poultry. It was impossible to ignore lupins since the crop had 32-45 % protein in seed and could be cultivated on light soils under temperate climatic conditions (Święcicki et al., 2015).

Compared to other legumes lupins are more tolerant to several abiotic stresses. The crop has the ability to recover poor and contaminated soils (Lucas et al., 2015) and withstand droughts (Fogelfors, 2015, p. 419). From the 1950s and forward the lupin production in Europe experienced a steadily decline. The contributed factors were low productivity due to seasonal variability, low price on lupin grain, and importation of soybean favored by EU policies. There has been an increase in production of lupin in Europe since 2003, which is due to re-established interest in legume crops (Lucas et al., 2015). In Sweden there is small-scale production of lupin beans for human consumption (NordiskRåvara., 2020) and cultivation of lupin for fodder production (Scandinavianseeds., 2020).

(12)

8 2.3.6 Quinoa

Quinoa (Chenopodium quinoa Willd.) is a pseudo-cereal (Manners et al., 2020). Quinoa originates from the Andean region where it has been cultivated for thousands of years. The crop produces seeds that varies in colors from white and yellow to purple and black. It is an adaptive crop that can be used in diverse agroecological conditions in the world. That was the reason why United Nations Food and Agriculture Organization (FAO) dedicated the year of 2013 to be the International Year of Quinoa (Ruiz et al., 2014). Quinoa is a close relative to fat hen (Chenopodium album) which was a secondary crop during the iron age in Europe. When quinoa was cultivated in Denmark it had a protein content of 12-16 %. Quinoa have large amounts of saponin in the seeds which is a bitter compound occurring in the seed hull and work as a defense against abiotic stresses such as downy mildew and pests, and is therefore attractive for organic production (Jacobsen, 2017).

Before consumption, the seeds require dehulling and extended washing because of the large quantity of saponins in the seed, that has a bitter taste. Seeds of the “sweet” low-saponin quinoa taste better than the bitter ones to animals such as birds and rodents. It is also more sensitive to

fungal diseases which affect the productivity (Ruiz et al., 2014). At the present quinoa is being tested in Northern Europe. The interest in quinoa is growing fast

around the world but there is still little production of quinoa outside of the Andes. To be able to successfully cultivate quinoa in Northern Europe some aspects must be considered. The only cultivars that could be grown are daylength neutral varieties grown in Europe. The cultivars are well adapted to sandy soils in Denmark but since quinoa is drought tolerant it could be possible to grow under rainfed conditions on lighter soils (Jacobsen, 2017). Quinoa can tolerate several different stresses such as salinity, cold air, high solar radiation, night subfreezing temperatures, and different soil pH values (Gonzales et al., 2009). Quinoa has relatively few difficulties with diseases and pests. Although the disease Downy mildew have been found everywhere quinoa is grown, and it thrives under humid conditions with temperatures of 15–20°C (Jacobsen, 2017). Quinoa is attractive for food and is currently produced in small-scale in Sweden (Nordisk Råvara., 2020a). It is possible to use quinoa as animal feed as well since quinoa has a high feed value (Jacobsen., 2017).

2.3.7 Soybean

The soybean (Glycine max) is sometimes referred to as an oil crop, but taxonomically it is a legume. It is one of the worlds most cultivated crop, and places after maize, wheat, and rice. Soybean is grown on all inhabited continents and it is therefore exposed to many different diseases and pests. The main diseases in Europe are downy mildew, bacterial blight, canker, and charcoal rot (Stoddard., 2017). Soybean is used for both food and feed and it has a protein content of 35-45 % (Adolfsson., 2013). For cultivation of soybean it is possible to use the same machinery as used for cereals, oilseed rape, and other legumes (Fogelberg and Recknagel, 2017).

Even though soybean has been cultivated for thousands of years, it is not until the 1700s that soybean is occurring in Europe (Fogelberg and Recknagel, 2017). Breeding of soybeans started in Sweden already in 1935 and four varieties were on the market. Uncertainty and low yields led to that the crop only were cultivated as a vegetable and not for fodder. Today there is no breeding in Sweden (Fogelfors., 2015, p. 517), but promising trial cultivation exists on Öland (Fogelberg., 2008). It is considered that under some circumstances it would be possible to grow soybeans up to 59°N (Fogelberg and Recknagel, 2017). Soybean is not being cultivated for production in Sweden, due to low demand on Swedish soy, the production industry has

(13)

9

difficulties of handling small deliveries directly from farms, and the soybean must first be refined to protein meal, and no such facilities exist in Sweden (Linnskog Rudh., 2018).

2.4 Flooding and drought in Sweden

In Sweden, changes in precipitation between periods 1991-2018 and 1961-1990 showed a yearly increase in precipitation for Sweden of 10 %. For winter, precipitation has increased up to 30 % in western parts of Sweden and the increase is smaller in eastern parts of Sweden with some decreases in some eastern parts. Spring is drier in the south and southeast Sweden, but in general Sweden is wetter, at some places above 10 %. For summer precipitation has increased considerably were changes sin precipitation varies up to 30 %. Small changes in precipitation were shown during fall in Sweden (SMHI., 2019). Precipitation in Sweden is expected to increase and in the end of the century yearly average precipitation is expected to be 20-60 % higher than for period 1961-1990. Precipitation is predicted to increase for all seasons, but most during winter (SMHI., 2020a). Climate simulations predict shorter periods of drought, approximately a day shorter than today (SMHI., 2017).

There are two types of flooding. The least severe is called waterlogging and that is when the root and parts of the crop are flooded. The more severe flooding is called a complete submergence and indicate that the entire crop is under water (VanToai et al., 2001). When the soil profile is water saturated for 1-3 days it indicates as short flooding. This kind of flooding do not cause big but limiting damages on soil and plants. A prolonged flooding is when the soil is water saturated for 1-2 weeks, which leads to lasting damages on soil and plants (Wesström et al., 2016). Flooding is identified as a major abiotic stress and the limitations it enforces on roots have noticeable effects on growth and development for crops. Flooding in spring can have great impact on seed germination and establishment (Parent et al., 2008). The longer the soil is water saturated the lower the yield is when harvest comes. Most crops can tolerate shorter periods of water saturation, but the tolerance varies depending on several factors, such as temperature, humidity, and the crop’s stage of development (Wesström et al., 2016).

A long period without precipitation is called drought. In Sweden, the longest period without precipitation yearly was 10-20 days between 1991 – 2013 (SMHI., 2017). Most cultivated crops have a decline in yield when impacted by drought. In countries were precipitation generally is sufficient for crops, water shortage may still occur over a few weeks period, which lead to substantial yield loss (Daryanto et al., 2015).

3 Materials and methods

Seven crops were selected for the analysis of suitable areas for current situation. The first step for that analysis was to define suitability indicators with suitability conditions for each crop regarding (i) soil texture were - clay separates were chosen as a variable, (ii) soil pH value, (iii) vegetation period length, (iv) and soil temperatures - calculated based on air temperatures. Geographic information system (GIS) was used for creating layers with all these suitable conditions for each crop. The created layers were overlayed with each other to get a single layer that showed where a crop is suitable for cultivation. Overlay is a method in GIS to identify areas were criterions are met (Esri., 2020). The layers with suitable conditions were further used to analyse current and future suitability for protein-rich crops in relation to the risk for wet and dry periods. For detailed description of the work carried out in GIS, see the thesis’s supplementary material.

(14)

10

3.1 Data

For this study different types of data were used (tab. 2). For an analysis of the current suitability for protein-rich crops, all variables in the table were included except for the wet spell and the dry spell – which were considered abiotic stressors and included further on in the analysis. The analyses on current and future suitability for protein-rich crops with risk for wet and dry periods uses the same data used for the first analysis on suitable distributions for protein-rich crops plus wet spell for wet periods or dry spell for dry periods.

Tab. 2. Data used in this study.

Data Name Format Resolution Source

Clay separates Dsms_ler_191205 Raster 50 m SLU, SGU

Soil pH value Ph_3006 Raster 10 km Jordbruksverket

Vegetation period t2m_nVegPeriod5 Vector 4 km SCID, SMHI

Wet spell Wetspell, rcp 4.5 mean, abs Vector 4 km SCID, SMHI

Dry spell Dryspell, rcp 4.5 mean, abs Vector 4 km SCID, SMHI

Monthly mean air temperatures

Air temperatures (month) Excel Meteorological observations, SMHI

3.1.1 Clay separates

The data for clay separates (percent clay) includes values for the topsoil and cover nearly all arable land in Sweden up to the county of Gävleborg. The data is based on interpolated sampling points, with an uncertainty that varies between regions and scales. For the entire map, standard error on clay separates is 5.6 percent and r²= 0.76. The map cover 3 million hectares of arable land and approximately 92 percent of the total areal of arable land in Sweden today (Söderström and Piikki., 2016).

3.1.2 Soil pH

Soil pH varies over Sweden and high pH values are present in regions with lime rich bedrock and soils rich in clay. Raster data on soil pH is based on soil samples from almost 12 600 sampling points (Paulsson et al., 2015).

3.1.3 Simulated Climate indicies

In this study three kinds of data were used from the climate index database for Sweden. Those are vegetation period, wet spell, and dry spell. The radiative forcing for the SCID are RCP 4.5 and RCP 8.5. For this study, the climate scenario RCP 4.5 were chosen for analysis of future periods. The reason for this is that Manners et al., (2020) found this scenario most probable and useful for future periods close in time (2050). This study refers to periods relatively close in time to their study and RCP 4.5 is based on the believeth that carbon dioxide emissions only will increase until the year of 2040 (SMHI., 2018). The climate scenarios for SCID is based on a group of global climate models and the regional model RCA4. The scenarios have calculated statistics that are based on results from every climate model separately. The result is then presented in average, minimum and maximum from the group of climate models. Every shapefile with geographical data contains several columns with data for time periods and seasons. For predefined time periods the results are presented as average and those periods are: P1 (1961-1990), P2 (1991-2013), P3 (2021-2050), and P4 (2069-2098). The seasons are either annual or divided after the four seasons containing three months for every season (SMHI., 2015).

(15)

11

The vegetation period is defined as the length in number of days from start to end of the vegetation period. The vegetation period is based on yearly annual data (SMHI., 2015), and for this study P2 were chosen since it is used on analyzing the current suitability for cultivating protein-rich crop and then this period is closest in time because it represent the most recent years. Since the vegetation period is to be combined with other data from recent years it is suitable to stay in the closest time period possible.

For the analysis on the risk for abiotic stresses in term of dry and wet periods both current (P2) and future time periods (P3 and P4) were chosen. The reason for this, is the possibility to compare the current risk with future risks. Both P3 and P4 were chosen in order to compare if the change is different between them. P1 was not chosen since there is no interest in this study to analyse the previous risks of abiotic stresses. The dry spell is defined as the maximal number of days in a row without precipitation (precipitation under 1 mm). This data is used in the study to indicate dry periods or drought. The wet spell is defined as the maximal number of wet days in a row with precipitation per day over 10 mm (SMHI., 2015), and used for an indication of flooding, waterlogging or water saturated soils.

3.1.4 Air temperature observations

Monthly mean air temperatures from 84 stations were retrieved from SMHI. Daily average is calculated from measured temperatures from stations (SMHI., 2020b). For this study monthly temperatures for March, April, May, and June for the years 2010-2019 was used in the calculation of soil temperatures and the average for the period 2010-2019 on the four months was calculated. An average air temperature for each month based on ten years.

3.2 Suitability conditions for protein-rich crops

The first step of the analysis for suitable areas for protein-rich crops under current situation was to create suitability indicators including conditions suitable for protein-rich crops. The factors of interest included in this analysis for each crop are suitable soils, suitable soil pH value, number of growing days, and minimum soil temperature for seed germination (tab. 3). In the subsequent analyses, which account for the current and future risk of wet and dry periods in suitable areas, another index with conditions indicating how each crop tolerates drought and flooding was made (tab. 4).

(16)

12

Tab. 3. Suitability indicators and corresponding conditions suitable for protein-rich crops.

Crop Soil Soil pH Vegetation period Soil temperature

Common bean (Phaseolus vulgaris L)

Sandy soils and clay up to 25 % (Fogelberg., 2008) 6.5-7.5 (Fogelberg., 2008). 125-150 days (Fogelberg., 2008) >10 ℃ (Fogelberg., 2008). Faba bean (Vicia faba) Clay with 15-60 % clay (Jordbruksverke t., 2013) 6–7 (Jordbruksv erket., 2013) 145 days (Jordbruksverket., 2013; Hagman et al., 2016). >4 ℃ (Jordbruksverket., 2013: Runåbergs fröer., 2020). Field Pea (Pisum sativum) Clay, 15–40 % clay. (Holstmark., 2016). >6 (Fogelfors., 2015; Olrog., 2004). 110–120 days (Olrog., 2004). No information Lentil (Lens culinaris) Lighter soils and clay with clay up to 25 % (Adolfsson., 2013). >6 (Adolfsson., 2013) 80 - 110 days (Adolfsson, 2013). >5 ℃ (Adolfsson., 2013). Narrow-leafed lupin (Lupinus angustifolia)

Sandy soils and clay, 0-25 % clay (Rahbeck Pedersen, 2004). 5-7 (Gresta et al., 2017). 90-150 days (Gresta et al., 2017). >4 ℃ (DLF-TRIFOLIUM A/S, 2015). Quinoa (Chenopodi um quinoa) Soils up to 60 % of clay (Adolfsson, 2013). >4.5 (Garcia et al., 2015) <150 days (Jacobsen, 2003). >4 ℃ (Adolfsson., 2013). Soybean (Glycine max (L.) Merr.)

Light soils and clay with clay up to 40 % (Adolfsson., 2013; Fogelberg., 2009) 6.5–7.0 (Fogelberg and Recknagel, 2017). 125-150 days (Adolfsson., 2013; Fogelfors, 2015). >10 ℃ (Fogelfors., 2015).

(17)

13

Tab. 4. Conditions for tolerance to drought and flooding/waterlogging for protein-rich crops.

Crop Drougth Flooding

Common bean (Phaseolus vulgaris L)

High sensitivity to drought (Daryanto et al., 2015).

Sensitive to flooding (White and Molano., 1994). Faba bean (Vicia faba) Sensitive to drought (Jordbruksverket., 2013) Affected by flooding 1-3 days. Dead by flooding 1-2 weeks (Enghag et al.,2016). Field Pea (Pisum sativum) Sensitive to drought (Fogelfors., 2015; Holstmark., 2016). Affected by waterlogging at vegetative stages. The least tolerant among the 7 grain legumes (Malik et al., 2015; Solaiman et al., 2007). Sensitive to flooding (Holstmark., 2016). Dies after 1-2 days of flooding (Enghag et al., 2016). Lentil

(Lens culinaris

Sensitive to drought and high temperatures during flowering and seed growth (Materne and Siddique., 2009).

Lentil does not tolerate waterlogging (Adolfsson., 2013; Materne and Siddique., 2009). Narrow-leafed lupin (Lupinus angustifolia) Tolerant to drought (Rahbeck Pedersen, 2004). Affected by waterlogging at vegetative stages. Intermediate tolerance among the 7 grain legumes (Malik et al., 2015;

Solaiman et al., 2007). Quinoa

(Chenopodium quinoa)

Tolerant to drought

(Jacobsen., 2017; Ruiz et al., 2014).

More affected by

waterlogging than drought (González et al., 2009). Soybean

(Glycine max (L.) Merr.)

Sensitive to drought (Fogelberg., 2009).

Soil waterlogging for 2 days give yield reduction of around 20 % (VanToai et al.,2001).

(18)

14 3.2.1 Suitable soils

To determine a crops suitable soil, the study focus on clay separate (percent clay) in the soil. Clay separate were chosen over soil texture even though both were available data from the same database. The reason for this is that soil texture in Sweden is divided differently from FAO, that is used in international contexts. The literature on the Swedish soil conditions for the crops were in Swedish and used the Swedish system for soil texture, and the conditional data had the international. In Eriksson et al., (2011) clay separate was found in relation to the soil texture. That relation was used in this study to transform soil texture into percentage of clay. For Sweden, clay data is divided after the percentage of clay, where for example sandy soils have low percentage of clay (Eriksson et al., 2011).

3.2.2 Suitable soil pH value

Both chemical, physical, and biological processes are affected by soil pH and it is therefore important to determine pH when a classification of the soil characteristics is in focus (Eriksson et al., 2011). Different crops have different pH and the pH value is of high importance for the variety of availability of plant nutrients. Soil pH varies over Sweden and high pH values are present in regions with lime rich bedrock and soils rich in clay (Paulsson et al., 2015).

3.2.3 Vegetation period

For this study, the data on vegetation period length was used to determine if the maturity days for a crop where within the length of the vegetation period. It is important that the vegetation period is long enough for the crops to mature since it affects the time for harvest and yield. Sweden have cold and humid autumns, which makes harvest more difficult because it leads to increasing drying cost and it reduces seed quality. Therefore, an early harvest is desirable. Late sowing of a crop and a cold growing season postpone growth and the crop must be harvested later (Jacobsen., 2017).

3.2.4 Soil temperature

Different crops need different soil temperatures for seeds to start growing. For seed germination to start, the minimum soil temperature for a crop needs to be reached. The soil is warmed up by heat from the sun, some amount of heat is also derived from the center of the earth, but that amount is of little importance in soil near the surface. The solar insolation reaches the ground directly or diffusely as well as indirectly through airflow, precipitation, and condensation of water vapor. The direct insolation varies with the tilt angle, south exposed soils are generally more warmed up than north exposed soils. Soil varies in their capacity to conduct heat which is determined by losing rate, air content, and water content. Soil conducts heat better if it has high water content than a dry content. Moreover, snow works as a heat insulator. Heat capacity, heat conductivity and evaporation work together to create prevailing soil temperatures under given climate conditions. During spring and early summer, clay– and peat soils are characterized as cold soils whereas gravelly soils and sandy soils are warm soils (Eriksson et al., 2011).

Measured soil temperatures would have been the best option for the analysis of current suitable soil temperature condition. However, for this study, soil temperatures were calculated from air temperatures since data with measured soil temperatures were unavailable. Hence, an alternative approach was needed. Jungqvist et al. (2014) have compared measured soil temperatures to model future soil temperatures on four locations in Sweden. For this study, an equation from one location were selected to calculate the soil temperature from air temperature. The location Gårdsjön was chosen since its middle soil layer has a depth of 10 cm which was closest to the topsoil and the location is within this study’s location. The equation made by

(19)

15

Jungqvist et al., (2014) shows the relation between air temperatures and soil temperatures. It is based on annual average and temperatures shifts around the year (Jungqvist et al., 2014). The equation for soil temperature on 10 cm depth, where Y represents soil temperature and X represents air temperatures:

γ=0.6679x+2.2498

The equation is used to derive the unknown soil temperature from the known air temperature. The data used for the calculations was monthly average air temperatures from 84 stations. This data was in excel documents. Temperatures varies from year to year, therefore averages of monthly average temperatures from 2010-2019 were made for 4 months (March, April, May, June) since it represents the months for spring sown crop cultivation. The air temperatures are measured 2 m above the ground.

3.3 Creating soil temperatures from air temperatures

In the GIS software Arc Map, an excel document containing 84 weather station with coordinate positions in Sweref 99 TM were transformed into a new workable table. X and Y columns were created in order to link them to the coordinate positions. The created XY table were interpolated into a point layer were the stations represent a point that is placed together with the geographical position. Now, 84 stations were placed on the right position in the coordinate system, but the stations lacked values for temperatures. Columns for March, April, May, and June were created in the attribute table. In those columns the average for ten years of monthly air temperatures were added. Later, this point layer was to be transformed to a raster layer, but those 84 points with data on air temperatures did not generate in a detailed raster layer. The soil pH raster layer that were being used later in the analysis had the lowest resolution of 10 km, and this layer with temperatures should not be lower than that since it affects the end result for the analysis. More point was created, using soil pH as a template. Those new points lacked information about temperatures, so they were given assumed temperature values after the surrounding points representing the stations that already had temperature values. Now, it was possible to transform the point layer to raster and using soil pH as a template without getting lower resolution than the soil pH layer and no spaces in the map that were not covered by any raster. A raster layer in 10 km resolution showing the average air temperatures over the last ten years for March, April, May, and June had been created. The average air temperature over ten years for March, can be seen further down (fig. 1).

For the analysis of suitable areas for protein-rich crops under current situation, the air temperatures needed to estimate the soil temperatures. Here, the equation from Gårdsjön were used. For every crop, a minimum soil temperature is known for germination to start and the equation is used to figure out what air temperature that represent. Raster calculator was used to estimate in which month the soil temperature reaches above a certain temperature. Values that reaches above a certain temperature get the value of 1 and temperatures that did not reach that temperature got the value 0. From the equation it was estimated that a soil temperature of 4℃ represent an air temperature of 3℃. Average air temperatures of 3℃ were found for the month of March, indicating that a crop that needs at least 4℃ in the soil can be sown in March and not earlier. Estimated soil temperatures above 4℃ for March (fig. 2) were the value 1 (dark blue) shows were the temperatures reached above 4℃ and value 0 (light blue) shows were temperatures did not reach above 4℃.

(20)

16

Fig. 1. Created raster layer, showing the average air temperatures (℃) in March 2010-2019.

(21)

17

3.4 Current suitable areas for protein-rich crops

For the analysis of suitable areas for protein-rich crops under current situation, the data on clay separate, soil pH, vegetation period length, and soil temperature were included in relation to the conditions needed for each crop. All data were in raster format, except for the vegetation period, which was in vector, so it was first transformed into raster. For each crop, raster layers were created with 1 and 0 values that were based on the suitability indicators for the suitability conditions for each crop. 1 represents values that reached a certain value, and 0 when a certain condition was not reached. This was made on clay separate, soil pH, vegetation period length, and soil temperatures separately, meaning that four raster layer per crop were created. These four layers showing were suitable conditions were reached were overlayed into one single layer for each of the crops. In the analysis, values for each fulfilled condition is calculated, but the interest lays in where all criterions for suitable areas are fulfilled. Therefore, a symbology was chosen that visually represent these areas only. Areas where calculated for suitable areas for each of the crops.

3.5 Current and future suitable areas for protein-rich crops with risk for wet and

dry periods

For this analysis, the created data based on suitability indicators and the corresponding conditions for each crop were used together with the conditions for tolerance to drought and flooding/waterlogging for protein-rich crops. The data on wet and dry periods are in vector so they were transformed into raster. For this analysis, the periods: 1991-2013, 2021-2050, and 2069-2098 where analysed for both wet and dry periods.

Fig. 3. Annual mean of wet days in a row with daily precipitation above 10mm for period:1991-2013, showing less than or equal to 2 wet days in a row and less than or equal to 3 wet days in a row.

(22)

18

Two wet days in a row was chosen for the analysis of risk for wet periods. One wet day is obtained in all of Sweden and less or equal to three days cover almost every part of Sweden. Two wet days are more interesting for the analysis, since two wet days shows more variation than one or three wet days (fig. 3). Since it is not really clarified how much rain a certain crop can survive, the same amount of days for all crops was used. The current suitable areas for the crops was overlayed with the <=2-days with high amount of rainfall. For every crop, a new layer with suitable areas with <=2 wet days were created. Areas were calculated to show the amount of area that is not exposed to flooding. This was done for three periods in order to find the suitable areas for crops not simulated to be exposed to flooding in the future.

Fig. 4. Dry days forperiod:1991-2013.

For the analysis of the risk for dry periods the start value were above 7 dry days. All areas in the study region were above 14 dry days, based on yearly average for the period: 1991-2013. Less than or equal to 21 dry days was chosen since many but not all areas had 21 days as a maximum value (fig. 4). Very few locations were above 28 days and they were not located within the study region. Since the interest was to identify suitable areas for different crops, less or equal to 21 dry days were chosen. The value that are not less than or equal to 21 days were areas with more than 21 days, but very few went over 28 days. The current suitable areas for the crops was overlayed with the <=21-days without precipitation. For every crop, a new layer with suitable areas with <=21 dry days were created. Areas were calculated to show the amount of area that is not exposed to drought. This was done for three periods in order to find the suitable areas for crops not stimulated to be exposed to drought in the future.

(23)

19

4 Results and discussion

The first question in this study was to locate were in Sweden the current suitable areas for cultivation of different protein-rich crops and how large those suitable areas are. The suitable areas and distribution for protein-rich crops are shown and discussed in the next section below. The second question is discussed in the last sections of the thesis - where in Sweden under current and future climates are the analysed suitable areas for protein-rich crops not exposed to wet and dry periods.

4.1 Current suitable areas and distributions for protein-rich crops

For all crops the first and second month to start sowing is March and April except for common bean and soybean were sowing is possible in May and June (fig. 5 and 6). It is the soil temperatures that determine when it is suitable for the seed to germinate, common bean and soybean needs higher soil temperatures than the other crops.

Quinoa is suitable on 99.9 % of the arable land in Sweden (tab. 5). This is probably due to the fact that quinoa tolerates all soil textures and had a clay percentage reaching up to 60 % and a soil pH that did not exclude any areas within the study (Adolfsson., 2013; Garcia et al., 2015). The soil temperatures did show when, and not where, it is possible for a crop to grow, so it does not explain why quinoa got this high suitability extent. The vegetation period was fulfilled for all crops within the study area, so it does not affect the result.

Regarding the soil temperature, the crops need to have time to mature before harvest takes place, which it needs to do before the vegetation period ends. The crops sown in March and April would have no problem with this. For common bean and soybean, the crop barely has time to mature before the vegetation period ends, but since the result with sowing in May and June correspond to the time of sowing in Swedish trials for soybean (Adolfsson., 2013; Fogelfors, 2015) it would be possible at least in the southern part of Sweden. Since soybean got high suitability in all parts of Sweden it is possible that the suitability in the middle of Sweden is more limited than this study shows -the areas that could be sown in May is probably more suitable than those in June, because the crop will have more time to develop before it must be harvested. It is interesting that the two crops sown later, both beans have so different results. Common bean is the crop with the least suitable areas and with the least percentage of total arable land (12.4 %). In contrast to soybean that has the most amount of suitable areas after quinoa and the suitable areas is 71.8 % of the total arable land in Sweden. Common bean is one of the crops that is currently being cultivated for production in small scale (Stoddard., 2017), while soybean is not, due to limitation in production possibilities (Linnskog Rudh., 2018). The change between them has to do with the soil texture and soil pH, where they differ. Soybean tolerates much higher percentage of clay (40 %), whilst common bean tolerates (25 %) of clay (Adolfsson., 2013; Fogelberg., 2008; Fogelberg., 2009). Common bean has suitable areas on Öland while soybean does not (fig. 6), this is related to the fact that the soil pH is slightly higher for common bean (6.5-7.5), than for soybean (6.5-7.0) and the suitability in Öland have to with the fact that the island have higher soil pH due to lime rich soils (Fogelberg., 2008; Fogelberg and Recknagel., 2017; Paulsson et al., 2015).

It is in the eastern part of middle Sweden that clay content with percentage of 25-60 is found (Paulsson et al., 2015), which is visible in this study where all the crops with a clay content over 25 % (faba bean, field pea, quinoa, and soybean) have suitable areas in that part of Sweden. Faba bean and field pea are two of the crops that are currently being cultivated (Stoddard., 2017). The result on their suitability are very similar in distribution patterns and the amount of suitable areas. Both crops have suitable areas in the most southern parts and the middle parts of

(24)

20

Sweden reaching from west to east. One exception is the field peas suitability in Gotland, that faba bean does not have. Field pea tolerates higher soil pH than faba bean (Fogelfors., 2015; Olrog., 2004; Jordbruksverket., 2013) and on Gotland there is lime rich soils that have higher soil pH values (Paulsson et al., 2015).

The current suitability for protein-rich crops in this study is based on four indicators, thus many factors that could affect suitability for crops are not included. Precipitation is for example not included in the current situation and would affect suitability and this is not taken in consideration here.

Table 5. Current suitable areas for protein-rich crops.

Current suitable areas for protein-rich crop

Suitable areas with start for sowing (ha)

% of farmland in Sweden 2015 1st_month ₁st_{+ 2}nd month % of arable land % of arable land & pastures

Common bean 290 000 320 000 12.4 10.5 Faba bean 160 000 1 310 000 50.6 43.1 Field pea 170 000 1 130 000 43.6 37.2 Lentil 100 000 1 420 000 54.8 46.7 Narrow-leafed lupin 450 000 1 530 000 59.0 50.3 Quinoa 510 000 2 590 000 99.9 85.2 Soybean 1 050 000 1 860 000 71.8 61.2

(25)

21

Fig. 5. Distribution of suitable areas for the crops that can be sown in March and April.

(26)

22

4.2 Future suitability for protein-rich crops in Sweden

Manners et al., (2020) is the only study found that compare many protein-rich crops over Europe in their suitability under both current and future conditions, with possibilities to compare results for Sweden. In this study, quinoa had the highest amount of suitable areas under current conditions. This is something that Manners et al., (2020) also found in their study, there 13 legumes and pseudo-cereals was analysed for their suitability in Europe. Common bean, narrow-leafed lupin, and chickpea was also found suitable under current conditions in their study. In this study chickpea was not analysed. According to Manners et al., (2020), the crop with the biggest suitability area for Europe and Sweden was quinoa, which was also found in this study.

In addition to quinoa, narrow-leafed lupin was one of the crops with highest suitability in this study, Manners et al., (2020) found the crop to be suitable on similar distribution and areas in Sweden for the future as the crop has today. This study found common bean suitable today and according to Manners et al., (2020), common bean will still be suitable in the future. In this study, common bean has the smallest suitable area. The suitability was smallest for common bean when the minimum temperature is reach in the study area. According to Fogelberg., (2008), the conditions needed to grow the Swedish brown beans which is belonging to common bean is the same for other beans as well such as kidney beans. That would mean that other varieties of common bean could be suitable under current conditions and in the future.

Faba bean, which does not belong to the other beans was not among those that Manners et al., (2020) found most suitable under current conditions or in the future. It is interesting that soybean was found to have large suitability in distribution and areas under current conditions in this study. Manners et al., (2020) did find soybean unsuitable for Sweden and most parts of Europe in the future. According to Manners et al., (2020), field pea was found to be more suitable for eastern Sweden and unsuitable in western Sweden in the future. The difference between east and west was not visible for this part of the study, which could have to do with the fact that the risk for wet and dry periods is not being considered. Manners et al., (2020) found lentil more suitable in southern locations and locations in height of the lakes Vättern and Vänern in future scenarios. Lentil under current conditions was found to be similar to Manners et al., (2020) in this study.

4.3 Current and future dry periods

Generally, there is an increase in suitable areas that are not exposed to the risk of three weeks of drought for all crops for period (P3) 2021-2050 and (P4) 2069-2098 compared to period (P2) 1991-2013 presented in (tab. 6) and visible in (fig. 7 and fig. 8). All the crops are also similar in their percentage of total suitable areas laying within areas with <=21 dry days for all three periods. Faba bean have the lowest value for P2 (71.0 %) and narrow-leafed lupin the highest (84.3 %). Common bean has the lowest value for P3 (93.8 %) and soybean the highest (97.8 %). Common bean has the lowest value even for P4 (90.6 %) and soybean the highest (96.2 %). For some crops there is a decrease in percentage between P3 and P4 (common bean, field pea, lentil, narrow-leafed lupin, and soybean), others show an increase (faba bean and quinoa). This is also represented in distributions of suitable areas where the increase for faba bean and quinoa for P4 is visible in purple (fig. 9). Distribution of suitable areas for the crops that are not exposed to the risk of three weeks of drought is spread out over Sweden (fig. 7 and fig. 8). Since the percentage of total suitable areas are high (71.0 % as lowest for faba bean in P2) suitable areas are similar to the suitable areas without risk for wet or dry periods.

(27)

23

From the conditions for tolerance to drought for protein-rich crops it was clear that all crops except for quinoa and lupin was sensitive to drought. Even though most of them are sensitive Daryanto et al., (2015) found that there were significant differences between different legumes in their adaptability to drought more specifically their ability to maintain high yield following a period of water stress.

Daryanto et al., (2015) found that faba bean had a yield reduction of 40 % when the water reduction was very high (>65 %). It was a global study, and Sweden is not seen as an arid or semi-arid country, but it still shows that faba bean is sensitive to water reduction whether Sweden reaches to this high level of water reduction or not. In this study, for P2 there is almost 30 % that is not located within <=21 dry days meaning that those areas would be more exposed to drought and higher yield losses. Those areas more exposed to drought, above 21 days, decreases for P3 and P4 to less than 10 %.

According to Daryanto et al., (2015), common bean had a 60 % yield reduction when the water reduction was 60-65 %. In this study, almost 20 % of the suitable areas have more than 21 dry days during P2. Those areas are more exposed to drought and higher yield losses than faba bean and at lower water reduction, less drought. Manners et al., (2020) saw an increasing suitability for common bean with 2 % from current condition to future condition 2050. That could be relating to this study, with more areas that are exposed to <= 21 days for P3 and P4 were the areas exposed to more than 21 dry days are less than 10 %. Common bean has high yield losses when exposed to drought, but it has also much less suitable areas than the other crops (tab. 6 and fig. 8). Daryanto et al., (2015) found that soybean had a yield reduction of 28 % when the water reduction was 60-65 %. Soybean is therefore much less sensitive to drought than common bean and faba bean. In this study, for P2, less than 20 % is exposed to more than 21 dry days, which is larger areas than common bean even though they have the same percentage since soybean is suitable on a much larger area than common bean, but soybean have a lower yield reduction at the same amount of water reduction. For P3 and P4 there are less than 5 % that is exposed to more than 21 dry days with higher risk of yield losses.

When the water reduction was >65 % lentil had a yield reduction of 22 % (Daryanto et al., 2015). Lentil needs to get exposed to higher amount of drought to get a yield reduction that is not so high. In this study, less than 20 % of the suitable area is exposed to more than 21 days of drought which is decreasing for P3 and P4 to around 5 %. Among the crops in this study it is one of the less sensitive to drought because it need higher water reduction and when exposed to drought it has a relatively low yield reduction, and a low percentage of the suitable areas are exposed to more than 21 dry days. According to Daryanto et al., (2015) field pea had a low yield reduction of 20 %, but with water reduction <60 %, which was lower than lentil. Field pea shows yield reduction sooner than lentil when exposed to drought. In this study, field pea has a higher percentage of suitable areas that are exposed to more than 21 dry days (27.4 %) than the other crops except for faba bean. Those areas exposed to more than 21 dry days decreases to around 5 % for P3 and P4, just as for many of the other crops in the study. Quinoa is considered to have high tolerance for drought and is chosen as a good candidate to be used for offering food security in a future scenario with increasing salinization and aridity in the world (Jacobsen., 2017; Ruiz et al., 2014). In studies where they reduced the water use up to 50 % on crops that had full irrigation, quinoa showed no effect on yield (Ruiz et al., 2014). This is probably because quinoa can increase its water use efficiency during soil drying (Jacobsen., 2017). Since quinoa is not previously being compared to the other chosen crops, it is harder to compare quinoa against the others. It is possible to think that quinoa is the crop in

(28)

24

this study that are most tolerant to drought. It is possible that the crop could endure the dry period of 21 days without any yield losses, the change of yield losses would be higher for areas exposed to more than 21 dry days. As discussed above those areas exposed to more than 21 dry days decreases considerably for P3 and P4.

Tab. 6. Suitable areas with <=21 dry days for the periods: (P2)1991-2013, (P3)2021-2050, and (P4)2069-2098.

<=21 dry days P2 (1991-2013) P3 (2021-2050) P4 (2069-2098)

Suitable areas for protein-rich crops Area (ha) % of total suitable area Area (ha) % of total suitable area Area (ha) % of total suitable area Common bean 260 000 81.3 300 000 93.8 290 000 90.6 Faba bean 930 000 71.0 1 230 000 93.9 1 250 000 95.4 Field pea 820 000 72.6 1 100 000 97.3 1 070 000 94.7 Lentil 1 160 000 81.7 1 360 000 95.8 1 340 000 94.4 Narrow-leafed lupin 1 290 000 84.3 1 490 000 97.4 1 460 000 95.4 Quinoa 2 080 000 80.3 2 460 000 95.0 2 470 000 95.4 Soybean 1 510 000 81.2 1 820 000 97.8 1 790 000 96.2

(29)

25

Fig. 7. Distribution of suitable areas for protein-rich crops sown in March and April with <=21 dry days in a row for the periods:(P2)1991-2013, (P3)2021-2050, and (P4)2069-2098. The map is presented with three overlapping layers. P2 is placed above the two other periods since it has the smallest area. Therefore, is all area for P2 present in the figure. When P4 is present on the map it is because it has suitable areas that were not suitable for P3 since P4 is placed below P3. When P4 is not visible it is related to the overlapping and then P3 and P4 share the same coverage.