Science of the Total Environment 822 (2022) 153514 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv Mapping direct N2O emissions from peri-urban agriculture: The case of the Metropolitan Area of Barcelona Angelica Mendoza Beltran a, Kelzy Jepsen a, Martí Rufí-Salís a, Sergi Ventura a, ⁎ Cristina Madrid Lopez a, Gara Villalba a,b, a Sostenipra Research Group (SGR 01412), Institute of Environmental Sciences and Technology (CEX2019-000940-M), Z Building, Universitat Autònoma de Barcelona (UAB), Campus UAB, Bellaterra, 08193 Barcelona, Spain b Department of Chemical, Biological and Environmental Engineering, Universitat Autònoma de Barcelona (UAB), Campus UAB, 08193 Bellaterra, Barcelona, Spain H I G H L I G H T S G R A P H I C A L A B S T R A C T • Spatially explicit direct N2O emissions maps from peri-urban agriculture are rare • There is lack of relevant datasets to esti- mate N2O emissions at peri-urban scale • We created maps of N2O emissions, harnessing bottom-up data of relevant factors • N2O emissions maps tackles spatial vari- ability but carry uncertainty of data used • N2O emissions maps highlight opportuni- ties for sustainable peri-urban agriculture A R T I C L E I N F O A B S T R A C T Article history: Geographically explicit datasets reflecting local management of crops are needed to help improve direct nitrous oxide Received 18 August 2021 (N2O) emission inventories. Yet, the lack of geographically explicit datasets of relevant factors influencing the emis- Received in revised form 24 January 2022 sions make it difficult to estimate them in such way. Particularly, for local peri-urban agriculture, spatially explicit Accepted 25 January 2022 datasets of crop type, fertilizer use, irrigation, and emission factors (EFs) are hard to find, yet necessary for evaluating Available online 29 January 2022 and promoting urban self-sufficiency, resilience, and circularity. We spatially distribute these factors for the peri-urban Editor: Damia Barcelo agriculture in the Metropolitan Area of Barcelona (AMB) and create N2O emissions maps using crop-specific EFs as well as Tier 1 IPCC EFs for comparison. Further, the role of the soil types is qualitatively assessed. When compared Keywords: to Tier 1 IPCC EFs, we find 15% more emissions (i.e. 7718 kg N2O-N year−1) than those estimated with the crop- Local food specific EFs (i.e. 6533 kg N2O-N year−1) for the entire AMB. Emissions for most rainfed crop areas like cereals (e.g. Urban agriculture oat and barley) and non-citric fruits (e.g. cherries and peaches), which cover 24% and 13% of AMB's peri-urban agri- GHG emissions maps cultural area respectively, are higher with Tier 1 EF. Conversely, crop-specific EFs estimate higher emissions for irri- Cities gated horticultural crops (e.g. tomato, artichoke) which cover 33% of AMB's peri-urban agricultural area and make GIS up 70% of the total N2O emissions (4588 kg N2O-N year−1 using crop-specific EFs). Mapping the emissions helps eval- uate spatial variability of key factors such as fertilizer use and irrigation of crops but carry uncertainties due to down- scaling regional data to represent urban level data gaps. It also highlighted core emitting areas. Further the usefulness of the outputs on mitigation, sustainability and circularity studies are briefly discussed. ⁎ Corresponding author at: Department of Chemical, Biological and Environmental Engineering, Universitat Autònoma de Barcelona (UAB), Campus UAB, 08193 Bellaterra, Barcelona, Spain. E-mail address: gara.villalba@uab.cat (G. Villalba). http://dx.doi.org/10.1016/j.scitotenv.2022.153514 0048-9697/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/). A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 1. Introduction Most N2O emissions estimates exist for the agricultural sector at global level (Tesfaye et al., 2021) or country level, as part of the National Inven- In the past century, agriculture has experienced an exponential growth tory Submissions made to the United Nations Climate Change Convention to meet the food demand of an increasing global population (Albanito et al., (UNFCCC) secretariat (UNFCCC, 2015, 1997). Fewer estimations of N2O 2017). This agricultural transformation has intensified agricultural systems emissions have been made for more localized agricultural areas and in con- and led towards the substantial use of nitrogen (N) fertilizer. As global food crete for peri-urban agricultural regions i.e. agriculture usually referring to demand is expected to increase, further pressure will be added on agricul- open-air, soil-based, ground-base, conventional agricultural practices tural soils by increasing N use (Muller et al., 2017). Yet, over half of the N around cities (Dorr et al., 2021) – definition also adopted in this study. fertilizer applied to agricultural soils globally is being lost, as only 47% of One of the reasons for this is the lack of finer-scale land cover datasets in- N applied is utilized to produce crops (Lassaletta et al., 2014). cluding relevant information. Most geographically explicit datasets include Direct nitrous oxide (N2O) is dissipated to the air when there are exces- aggregated typologies of croplands and lack information on agricultural sive N inputs to agricultural soils, which then influence microbial processes management practices at relevant resolutions, which are vital to answer of nitrification and denitrification. Nitrification occurs in aerobic condi- questions related to biogeochemical cycles (Monfreda et al., 2008), re- tions and can be described as the oxidation of ammonium to nitrate. Deni- source circularity and for more accurate greenhouse gas estimations. More- trification occurs in anaerobic conditions and can be described as the over, peri-urban agricultural areas are becoming increasingly relevant reduction of nitrate to molecular N (Leip et al., 2011). N2O is a greenhouse particularly in the light of pandemics and the most recent COVID-19 crisis gas (GHG) that contributes to ozone depletion and climate change. At a (Vittuari et al., 2021), not to mention the imminent climate crisis (IPCC, global level, croplands are the main anthropogenic source of direct N2O 2021) as they may increase resilience and sustainability and self- emissions (Wang et al., 2020). Global cropland N2O emissions estimated sufficiency of cities' food supply (Langemeyer et al., 2021). Assessing the by bottom-up approaches ranged between 1.5 and 5.0 Tg N year−1 benefits and costs provided to cities by peri-urban agricultural areas re- (Wang et al., 2020). quires appropriate geographically explicit datasets, at relevant resolutions There are several ways to estimate direct N2O emissions. The use of and with relevant information. emission factors (EFs) is among the most common ones. An N2O EF can In this study we aim to provide a method for more accurate estimation be defined as the mean emission rate per unit of N applied (kg N2O-N/kg of direct N2O emissions, building on the creation of a peri-urban agriculture N applied) (De Klein et al., 2006; Yue et al., 2019). The Intergovernmental map including irrigation and fertilization information, as well as crop- Panel on Climate Change (IPCC) provides broadly used guidelines for esti- specific EFs, at an appropriate resolution. We harness the best available mating direct N2O emissions from agricultural soils. The “Tier 1” methodol- data and combine land cover maps, local agricultural statistics (e.g. individ- ogy uses a default EF of 0.01 kg N2O-N/kg N, assuming that 1% of N applied ual crop type, crop surfaces), management practices (e.g. kg N fertilizer and to agricultural soils, in the form of mineral and organic N applications, in- irrigation applied), a soil map and EFs (i.e. IPCC Tier 1 EF and crops-specific cluding crop residues and mineralization of soil organic carbon, will be re- EFs) to obtain a novel peri-urban agriculture map and spatially distributed leased into the atmosphere as N2O (De Klein et al., 2006). Nonetheless, this emissions. We focus in the peri-urban agricultural sector of the Metropoli- EF overlooks other key factors when determining N2O direct emissions. tan Area of Barcelona (AMB for its acronym in Catalan) in Spain where In fact, several studies (Bouwman et al., 2002; Flynn et al., 2005; Leip such map and N2O emissions distribution is not existent, as known to us. et al., 2011; Shepherd et al., 2015; Yue et al., 2019) and the IPCC as well We further briefly discuss the usefulness of the outputs on mitigation, sus- (De Klein et al., 2006), indicate that N2O EF's are influenced by character- tainability and circularity studies, among others. istics that are unique to a certain crop and region and that emissions vary spatially and temporally due the differentiation of local management prac- 2. Methodology tices and climate conditions (Gerber et al., 2016; Guevara et al., 2019; Wang et al., 2011). Other factors that are dependent on the crop type are Four basic steps were followed to spatially distribute the direct N2O the N fertilizer inputs and irrigation applied (Aguilera et al., 2013). Accord- emissions of peri-urban agriculture (Fig. 1). First, the study area was de- ing to Brentrup et al. (2000) it is estimated that almost 80% of N2O emis- fined and data was collected for it, in particular, individual crop manage- sions from agricultural soils are linked to mineral and organic fertilizers. ment data, and crop-specific emission factors. Second, cropping systems, Fertilizers provide available N to the soil for denitrification and nitrification defined as a single crop with its specific irrigation and fertilizer manage- processes. ment, were selected based on the available data. Third, two approaches Moreover, studies emphasize the importance of precipitation, tempera- were used to calculate direct N2O emissions per crop, 1) the IPCC Tier 1 ture, and soil climatic factors for the activation of nitrification and denitri- method (Tier 1) and 2) using crop-specific EFs for selected cropping sys- fication (Corre et al., 1999; Dachraoui and Sombrero, 2020; Dobbie et al., tems (Tier 2). Finally, calculated emissions are mapped. In this section we 1999). Rochette et al. (2008) explains how precipitation and high temper- describe each step. atures enhance the soil moisture content, which then increases the water filled pore space in agricultural soils. As moisture fills the soil pores, there 2.1. Data collection for cropping systems and study area is less available oxygen in the soil, promoting conditions for denitrification. Thus, establishing regional EFs that consider these factors can improve di- The study area for this work is the Metropolitan Area of Barcelona rect N2O emission estimations. The IPCC indicates that a “tier 2” method (AMB). AMB is in the North-East region of the Iberian Peninsula and it in- to calculate direct N2O emissions should use specific EFs and these corre- cludes 36 municipalities of the autonomous community of Catalonia in spond to regional EFs for specific conditions (climatic and of cultivation) Spain (see Supporting information (SI) 1, Fig. S.1). Peri-urban and urban and N inputs (e.g. for mineral and for organic fertilizers) (De Klein et al., agriculture are widespread agricultural practices in AMB, although they 2006). have decreased in the last 50 years (Rufí-Salís et al., 2019). Scenarios of STEP 1 Method 1: STEP 2 Tier 1 IPCC EF Data STEP 3 STEP 4 Selection of collection for Calculation of Mapping N2O cropping cropping N2O emissions Method 2: emissions systems systems Crop-specific EF Fig. 1. Methodological steps followed in this study to map peri-urban agricultural direct N2O emissions for the Metropolitan Area of Barcelona. 2 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 urban and peri-urban agricultural expansion within the AMB and its sus- For the crop area values in Eq. (1), a collection of cartographies from the tainability challenges are explored in the Land Use Master Plan (PDU for AMB were hybridized (see Section 2.4). Together, fertilizer rate and crop its acronym in Catalan) which has green infrastructure as one of the 6 key areas, yield a map for annual N fertilizer application per crop (Fig. 3). themes currently under development (Barcelona Regional and AMB-PDU, 2019; Metropolitan Area of Barcelona, 2021). At present, the AMB encom- 2.3. Calculation of N2O emissions for selected cropping systems passes 5584 ha of cropland (Fig. 2). The year under assessment is 2015, as it is the year with the most recent Land Cover Map of Catalonia (CREAF, We use two methods to estimate direct N2O emissions according to 2015). This map, referred to as the MCSC map, is central to the crops' Eq. (2). selection and emissions mapping, as will be further described. It contains  a high-resolution thematic cartography of the main types of land cover of Direct N2 O Emissions N2 O−N year−1 the country (forests, crops, urban areas, etc.). Moreover, individual crops ¼ Emission Factor ðkg N2 O−N=kg NÞ  are not included in this map, but are described at the level of herbaceous  Annual N fertilizer application per crop kg N year−1 (2) crops, fruits, vineyards, olives, vegetables and greenhouses among other aggregated crop-groups or land use categories. The methods vary the EFs (Tier 1 and Tier2) and use the map of fertil- During the initial collection of data, individual crop management data, izer application per crop. and EFs were reviewed. For crop management, the review revealed fewer data at the AMB level, and most available sources reported data at more ag- 2.3.1. Tier 1: IPCC emission factor gregated regional scales e.g. country (Spain), autonomous community (Cat- Method 1 follows the IPCC's tier 1 guidelines for estimating direct N2O alonia) or provincial (Barcelona) levels. Therefore, data at the Catalonia emissions from agricultural soils. IPCC's Tier 1 emission factor assumes that and Barcelona level was collected. For each level, we recorded the crop 1% of the N applied into the soil will be emitted as direct N2O emissions (De types either for individual crops (e.g. wheat) or for crops groups (e.g. ce- Klein et al., 2006). Therefore, in Eq. (2), the EF is replaced for 0,01 kg N2O- reals) - depending on the availability, their production surface (ha) and N/kg N for all crops and this is multiplied by the estimated annual N fertil- yield (kg/ha) (Generalitat de Catalunya, 2016; MAPA, 2016). Moreover, izer application per crop to estimate the direct N2O emissions. The emis- their irrigation type (whether irrigated or rainfed) and annual fertilizer ap- sions calculated from this method are used as a reference for comparison. plication (kg N year−1) were also recorded (García-Serrano Jiménez et al., 2010; MAPA, 2016; Mercade et al., 2010). 2.3.2. Tier 2: crop-specific emission factors Regarding the EFs, the study by Cayuela et al. (2017) is a key source em- EFs for this method are adapted from the Cayuela et al. (2017) Mediter- ployed. We use a subset of its meta-analysis (see Section 2.3.2 for details on ranean meta-analysis. To fit the area of this study, the meta-analysis was re- this calculation) of available measurement data of direct N2O emissions fined by excluding case studies of used inhibitors and rice crops, because from Mediterranean cropping systems and the detailed EFs at the individual they are not representative for the AMB, and by adding cases from Pi and crop level to further calculate the crop-specific EFs. Deng (2019), that provide additional N2O estimations for cropping systems of interest in the AMB, particularly the Fruit crops. As a result, 22 case stud- 2.2. Local activity data ies and 93 different N2O EFs were included, 12 of these case studies are from Spain, and the rest are from Australia, Chile, Italy, and the USA. Local activity data corresponds to total annual N fertilizer application Using SPSS statistical analysis software, new mean EFs are established estimates per crop (kg N year−1). To calculate these values, as on-site based on the following conditions: crop type, irrigation type (irrigated or data was not available, we used a bottom-up approach based on the N fer- rainfed) and fertilizer type (mineral or organic) and fertilizer application tilizer rate per crop (kg N ha−1) and area of each crop plot (see Section 2.4 rate (kg N ha−1). We refer to these EFs as crop-specific EFs (SI 1, for the calculation of these areas), following Eq. (1). Table S.2), also shown in Table 1 per crop group. The SI 2 supplies the meta-analysis subset where crop-specific EFs as well as crop group EFs  are derived. The latter were applied to individual crops lacking crop- Annual N fertilizer application per crop kg N year−1  specific EFs for the area of study. For example, the meta-analysis provided ¼ Area of individual crop ha year−1  EFs for oat, however, because the rainfed EFs in the meta-analysis only has  N fertilizer rate per crop kg N ha−1 (1) one literature source, the specific crop EF based on this source was not used. Instead, the cereal crop group of 0.0029 kg N2O-N/kg N was applied. Three sources of data were used for the N fertilizer rate per crop in Eq. (1): 1) The N requirements per crop reported in the “Practical guide to 2.4. Mapping N2O emissions the rational use of fertilizers on crops in Spain” by the Ministry of Agriculture, Fisheries and Food (García-Serrano Jiménez et al., 2010) assuming total N 2.4.1. Crop areas fertilizer application to crops following the crops' requirement of N and is This section describes the methodology for creating the crop map for thus applied in an optimal way; 2) the estimation of N fertilizer rate per our study region. For this purpose, two maps have been used. First, the crop from the HERMESv3, a multiscale atmospheric emission modelling DUN-SIGPAC (as we refer to it in this study) is a map provided by the De- framework (Guevara et al., 2019), for what the model uses data from the partment of Agriculture, Livestock, Fisheries and Food from the Generalitat Spanish Ministry of Agriculture, Fisheries and Food on surfaces and yield de Catalunya (DARPA, 2015) containing the annual declaration of crops (MAPA, 2015) and the “Practical guide to the rational use of fertilizers on (Declaració agrària, DUN) made by farmers in Catalonia, which matches crops in Spain” (García-Serrano Jiménez et al., 2010), as well as from the public registry of Geographic Information System of agricultural parcels Mueller et al. (2012); and 3) for crops not available in the two previous in Catalonia (Sistema d'informació geogràfica de parcel·les agrícoles, sources, the N fertilizer rate was calculated from the initially collected SIGPAC for its acronym in Catalan). This map contains very detailed but in- data at the Catalonia level. For the latter, we use the total annual production complete information on the province, district, municipality, crop-group, and annual surface of the crop to calculate an average annual yield in ton individual crops and whether they are irrigated, on an annual basis. Second, per ha (Generalitat de Catalunya, 2016). Then, the coefficients of N extrac- the MCSC map (CREAF, 2015) contains less detail on agricultural parcels, tion in kg N per ton, was derived from the Nitrogen Balance from the Span- but has more spatially complete information. ish Government (MAPA, 2016) and by means of these two values the These two maps have been hybridized into one containing the detail of average fertilizer rate per ha was derived (SI 1, Table S.1). Table S.2 in SI DUN-SIGPAC map and the spatial completeness of MCSC map, by means of 1 shows the values and data sources used for the N fertilizer rate per the generalization of the DUN-SIGPAC crop information as follows. The crop, also shown per crop group in Table 1. first step was to apply the Voronoi diagram or Thyssen polygons (Gold 3 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 4 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 and Angel, 2006) on the DUN-SIGPAG polygons. This methodology divides 2.4.3. Qualitative component: soil map the plane into regions represented by a polygon with a center point for each The effect of soil characteristics on EFs was not quantified. We, how- region. For each center point the closest region values are assigned. This ever, add a soil map layer with qualitative information to as why emissions procedure generates the Voronoi polygons, with DUN-SIGPAC information are higher in certain areas compared to others. The Institut Cartografic i associated. The next step is the validation of the associated information to Geologic de Catalunya (ICGC) provides the Catalonia soil map (Mapa de each of the polygons by means of a Python code. This code contains a dic- sòls de Catalunya) (ICGC, 2018), that uses the international World Refer- tionary that relates the crop categories of both maps, having more specific ence Base classification system (see Fig. S.5). This layer enables the descrip- categories for the DUN-SIGPAC map with respect to those of the MCSC. For tion of influencing factors such as climatic factors on quantified emissions instance, the typology Citrus Fruits in MCSC map relates to Lemon, Manda- in the AMB. Most of the peri-urban agriculture in the AMB fall within rin, Orange and other Citrus in the DUN-SIGPAC map typologies. The code Calcaric Fluvisol, Fluvic Arenosol and Eutri Leptosol soils. checks if there is a match between DUN-SIGPAC and MCSC land cover ty- pologies. When the two categories coincide i.e. belong to similar descrip- 3. Results tions e.g. Mandarine for DUN-SIGPAC and Citrus Fruits for MCSC, a positive result is obtained. We have positive results in about 40% of the This section displays the results of each step completed to calculate di- polygons, where the DUN-SIGPAC information is used. To avoid the mis- rect N2O emission maps. match of the 60% remaining i.e. where the more detailed DUN-SIGPAC First, Fig. 3 panel A, shows the raster map for the annual N fertilizer ap- and more generic MCSC crop typologies do not match (due to differences plications at a 100 × 100 meter resolution for the AMB. Fig. 3 panel B, in the source of information) the more generic typology associated with shows the vector map with the crop-specific EFs per parcel on the URBAG the MCSC map is used. map. Data from two sources has been used for the irrigation method per crop and polygon i.e. irrigated or rainfed. First, cadastral cartography data 3.1. Tier 1 N2O emissions (Dirección General del catastro, 2014) has been integrated onto the MCSC map repeating the Voronoi polygons methodology previously de- The IPCC Tier 1 EF assumes that 1% of the fertilizer applied to a crop is scribed, assigning irrigation/no irrigation to each polygon. The second released to the air as N2O. Thus, emissions in Fig. 4 (left map) correlate lin- source is the MCSC map itself that classifies land under irrigated or non- early with the N fertilizer input i.e. emissions are higher for parcels with irrigated. For both sources, a similar number of polygons are irrigated i.e. higher N fertilizer inputs and vice versa. Because this method does not rec- 52.6% according to Cadastral data and 54,25% according to MCSC map. ognize local climatic factors, management practices nor crop type, emis- Finally, information about the municipality in which each MCSC map sions produced in this map are a rough estimation for the AMB and are polygon is located is added to the resulting map. To do this, municipal meant to serve as a first reference level of data for this study. Total direct boundaries (Institut Cartogràfic i Geològic de Catalunya, 2020) have been N2O emissions for peri-urban agriculture in the AMB based on the Tier 1 used. A division has been applied to polygons situated in more than one mu- method are 7718 kg N2O-N year−1 i.e. 12.12 tons of N2O year−1, which nicipality, resulting in a map with more derived polygons but with munic- using a global warming potential of 298 kg CO2eq per kg N2O according ipal information for each of them. to the IPCC Fifth Assessment Report leads to 3611 tons CO2eq year−1. A more detailed explanation of the workflow used to generate the hy- brid map is provided in the SI 1 (Fig. S.2). To summarize, the resulting vec- 3.2. Tier 2: crop-specific emission factors tor map (from now on referred to as the URBAG AMB map since it is fruit of the ERC-Consolidator project Integrated System Analysis of Urban Vegeta- Fig. 4 (right map) shows the direct emissions considering the Tier 2 tion and Agriculture, URBAG) includes the following information associ- crop-specific EFs. Because the EFs come from studies applying similar agri- ated to each polygon: the crop type according to the Voronoi cultural practices, for example, whether crops are rainfed or irrigated, and methodology and further dictionary check (Fig. 2), irrigation according to uses both mineral and organic fertilizer, these emissions are expected to be MCSC and Cadaster data (SI 1, Figs. S.3, S.4) and its municipality (SI 1, more accurate than those estimated under Tier 1. Total direct N2O emis- Fig. S.1). From this map we derive the area per crop used in Eq. (1) to deter- sions for peri-urban agriculture in the AMB based on the Tier 2 method mine the total fertilizer input per crop area (the individual polygons). are 6533 kg N2O-N year−1 i.e. 10.25 tons of N2O year−1, which using a The URBAG AMB map also provides data on the irrigation method used global warming potential of 298 kg CO2eq per kg N2O according to the per crop area, in order to assign the right crop-specific EFs when applying IPCC Fifth Assessment Report leads to 3056 tons CO2eq year−1. Tier 2 method to estimate direct N2O emissions, for which the MCSC irriga- tion data is used. The largest crop groups in the URBAG AMB map, includ- 4. Discussion ing irrigated and rainfed areas, are the cereals, vegetables and non-citric fruits with a share of 36, 33 and 23% respectively, of the total peri-urban 4.1. Differences between Tier 1 and Tier 2 methods areas in the AMB (Table 1). The map attributes are described in the SI 1, and the shape file is also provided in SI 3. Overall, the IPCC's default value of 0.01 kg N2O-N/kg N is higher than the crop-specific EFs for some crops e.g. Sorghum, Barley, Wheat, Oat, Trit- 2.4.2. Calculation of spatial emissions for Tier 1 and 2 methods icale, Olive, and Almond and lower for others e.g. Cucumber, Tomato, and The local activity data was also integrated into the attribute table of the Vineyard crops. The lower values of the crop-specific EFs, especially for URBAG map. Specifically, the annual N fertilizer application per crop in ki- rainfed crops, leads to lower emissions compared with IPCC tier 1 default lograms, according to Eq. (1), the IPCC Tier 1 EF, and the crop-specific EFs EF estimates and the higher values of the crop-specific EFs, for irrigated were integrated for each polygon. This creates more detailed attribute ta- crops, leads to higher emissions, compared with IPCC tier 1 default EF bles which enables the calculation of direct N2O emissions at a polygon estimates. level using Eq. (2). As a result, two direct N2O emission maps are created Some selected examples for key crops illustrate these observations. For and compared. One map shows the IPCC tier 1 emissions, and the second instance oats encompass 249 ha in the AMB and emit 292 kg N2O-N year−1 shows crop-specific emissions. as estimated with Tier 1 CF while using the crop-specific EF lead to 94 kg Fig. 2. Crop areas in the Metropolitan Area of Barcelona according to the URBAG map, used as the base for determining N2O emissions. Total areas match the 2015 Land Cover Map of Catalonia (MCSC) by the Centre de Recerca Ecològica i Aplicacions Forestals (CREAF, 2015) and added detailed crop areas correspond to DUN-SIGPAC map (DARPA, 2015). 5 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 Table 1 Catalonia or Barcelona province. This meant that using these data or down- Crop group areas according to the URBAG map, fertilizer rate and emission factors. scaled data to represent the AMB level practices, introduced uncertainty Crop group Crop group Area Share Emission factor (kg Fertilizer rate due to geographic representativeness. Accuracy is lost when mapping the (Catalan) (English) (ha) N2O-N/kg N) (kg N ha−1) N fertilizer rate per crop. Crop-specific CFs, are based on field data, for Cereals Cereals 2024 36% 0,0029 167,9 which accuracy is expected to be better. However, despite of data being rel- Horticoles Vegetables 1868 33% 0,0148 165,7 evant for the AMB peri-urban agriculture systems, regions outside of the Fruita Dolça Non-Citric fruits 1272 23% 0,0038 61,8 AMB are included. Therefore, some uncertainty remains. Vinya Vineyard 131 2,3% 0,0152 59,6 Such uncertainties call for caution in the use of the N fertilizer rate maps Guaret Fallow Land 80 1,4% 0 0 Oliverar Olives 56 1,0% 0,001 57,6 and N2O emission maps i.e. values are indicative, until calibration and ver- Citrics Citrics 3 0,05% 0,0028 –a ification are undertaken. We nonetheless, believe the developed maps give Altres Usos Other uses 150 2,7% SI 1, Table S.2 SI 1, Table S.2 the order of magnitude of what can be expected on the field including key Total Total 5584 100% – – variability aspects and despite geographic uncertainty of some activity a No citric crop group in the URBAG map. data, they are valuable for other modelling purposes (see Section 4.4). Fu- ture work should focus on improving data gaps by conducting on-site sur- veys with local farmers and applying other participatory approaches on N2O-N year−1. Also, for all other cereal crops N2O emissions are larger for key peri-urban agriculture core areas or gathering field measurements for Tier 1 than for Tier 2 e.g. Barley (Tier 1: 568 kg N2O-N year−1; Tier 2: emissions were possible. Given the large variability in crop management 160 kg N2O-N year−1), Wheat (Tier 1: 103 kg N2O-N year−1; Tier 2: practices, capturing it with field campaigns requires careful design of the 34.2 kg N2O-N year−1), and Triticale (Tier 1: 125.2 kg N2O-N year−1; campaigns e.g. including measurements over a long enough period of Tier 2: 36.3 N2O-N year−1), as well as Almond (Tier 1: 5.4 kg N2O-N time such a year. Ideally the whole N cycle could be monitored opening a year−1; Tier 2: 1.7 kg N2O-N year−1) and Olive perennial crops (Tier 1: window to the assessment of all N-related inputs e.g. including organic fer- 32.4 kg N2O-N year−1; Tier 2: 3.2 kg N2O-N year−1). tilizers and outputs e.g. including other N-related emissions, as well as op- Horticultural crops makeup 1868 ha in the AMB – all irrigated, and is portunities to create a more circular peri-urban agriculture sector. Besides, the second largest land cover group. IPCC tier 1 method leads to 3098 kg such on-site data could help to assess to what extent downscaling or using N2O-N year−1, while crop specific EFs leads to 4588 kg N2O-N year−1. Veg- data for larger geographic scales introduces uncertainty to the estimated etables represent 40% of total emissions estimated with Tier 1 CF and 70% emission inventories. of total emissions estimated with crop-specific CFs (See Table S.3 and Fig. S.6 for results per crop group). 4.3. The role of crop management and soil types in N2O emissions Direct N2O emissions from irrigated vineyard crops are also lower ac- cording to the IPCC Tier 1 EF (Tier 1: 0.56 kg N2O-N year−1; Tier 2: Following the World Reference Base International Soils Classification, 0.86 kg N2O-N year−1). According to the 2016 Spanish Nitrogen Balance, the highest direct N2O from horticulture crops in the AMB came from for the year of 2016 Vineyards only received organic fertilization (MAPA, Calcaric Fluvisol i Fluvic Arenosol soils. According to the Soil Atlas of 2016) and according to the “Practical guide to the rational use of fertilizers Europe, Calcaric soils are composed from calcium carbonate build up and on crops in Spain” (García-Serrano Jiménez et al., 2010), the recommended cover many areas in Spain (JRC, 2005). Fluvisols are also common soils in fertilization rate is below 100 kg N ha−1 i.e. according to our own calcula- European Mediterranean regions and are often areas growing vegetables. tion 59.6 kg N ha−1. Applying a crop-specific EF that considers this fertili- This type of soil is often found near rivers, for example, in alluvial basins zation practice, results in higher emissions compared to the IPCC Tier 1 EF. of Spain such as the Baix Llobregat, the region with the most area of horti- Fig. 5 shows the difference between the emissions per year estimated culture within the AMB (see Fig. S.5). Collectively, horticultural crops pro- with tier 1 and tier 2 emissions factors (see Table S.3 and Fig. S.6, for results duced the highest emissions in the area of study according to Tier 2. For the per crop group). Tier 1 leads to 15% more emissions than Tier 2. The gen- year of 2015, our estimates show the AMB emitted 3098 kg N2O-N year−1 eral trend shows larger emissions estimated for irrigated horticulture crops (via Tier 1) and 4587 kg N2O-N year−1 (via Tier 2) from horticultural crops and rainfed vineyards with Tier 2 EFs, while for rainfed herbaceous crops, growth. Ultimately, higher direct N2O emissions could be linked to denitri- olives and non-citric fruits fewer emissions are estimated with these EFs fication due to intense management practices and climate characteristics in in comparison with IPCC tier 1 EF. Baix Llobregat. In the summer, Mediterranean agricultural systems experi- ence intensified management practices due to warmer temperatures and a 4.2. Mapping emissions: variability and uncertainty factors lack of rain during this time (Aguilera et al., 2013). According to the Idescat (2015) meteorology report, in 2015 the Baix Llobregat region expe- Direct N2O emissions are sensitive to local crop management practices rienced an annual average temperature of 16.6 °C and an annual mean rain- of irrigation and fertilization, among the most important, and climatic fac- fall of 276.6 mm for the year. In terms of irrigation practices, the Baix tors, which vary both spatially and temporally. Spatially resolving the emis- Llobregat region contains the greatest number of irrigated crops in the sions with the help of GIS, aids to account for key spatial variability factors. AMB. Lettuce and artichoke, typical crops grown in the Agrarian Park of For instance, by including site-specific characteristics of irrigation and fer- Baix Llobregat, use furrow irrigation as a common practice, and have tilization according to the crops grown in the AMB, the bottom-up approach been found to be over irrigated (Aljoumani, 2012). Similarly, Serra et al. used in this study helps to display the spatial variability of N2O emissions (2016) found poor efficient irrigation practices in the Agrarian park of due to these factors. Also, a regional more accurate estimation of direct Baix Llobregat, with only 26.4% of water applied contributing towards irri- N2O emissions is possible by making geographically explicit such variable gated crops. Furthermore, “Estanyar” is a traditional agricultural practice in factors. the Agrarian park of Baix Llobregat, where fields are flooded to wash away Regarding temporal variability, crop management and climatic factors salt in the soil (Serra et al., 2018). Generally, using intensive irrigation/ are also variable. Seasonal changes of these factors should be considered. flooding practices increases the moisture content in agricultural soils and We used the average yearly input of fertilizers and the irrigation method thus promotes the formation of N2O emissions. Regarding fertilizer but the seasonal fertilizers and water used e.g. along a year, could be management, horticulture had the second greatest fertilizer inputs as more accurate and tackle temporal variability of emissions. Such temporal estimated in this study, totaling 310 tons N year−1, after cereals with considerations can be accounted for in a geographically explicit manner 324 tons N year−1 as input. Mediterranean agricultural soils have experi- too, and added on top of the spatial considerations already included here. enced an increased amount of over-fertilization (Peñuelas et al., 2013) There are also large uncertainties around the results presented here too. where high soil nitrate near intensive agricultural areas have been linked Activity data was obtained from data at aggregated geographic levels of to chemical fertilizers, as well as denitrification linked to nitrate from 6 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 A. B. Fig. 3. A) Annual N fertilizer application for peri-urban agricultural in the Metropolitan Area of Barcelona, raster map 100 × 100 meter resolution. B) Crop-specific emission factors for peri-urban agriculture in the Metropolitan Area of Barcelona as estimated based on Cayuela et al. (2017). 7 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 Fig. 4. Direct N2O-N emissions estimated with the Tier 1 (left map) and Tier 2 (right map) emission factors for peri-urban agriculture in the Metropolitan Area of Barcelona, raster maps 100 × 100 meter resolution. organic and inorganic fertilizers (Postigo et al., 2021). High emissions of sustainability and circularity strategies. For instance, overly irrigated or fer- horticultural crops in Baix Llobregat could be related to the high N inputs tilized peri-urban agriculture areas will benefit from strategies such as crop leading to nitrate in the soil and the high irrigation inputs limiting oxygen diversification, reduction of overfertilization, optimal water management, availability due to water flowing into the soil pores instead. Thus, under alternative fertilization methods - such as inoculants (Arcas-Pilz et al., wet soils, and warm temperatures (16.6 °C) anaerobic conditions are pro- 2021), nutrients recirculation (Rufí-Salís et al., 2020), etc. Moreover, spa- moted. This activates microbial activity for denitrification. More specifi- tially distributing all nitrogen flows i.e. not only N2O emissions, is essential cally, the microbes breakdown the nitrate in the soil and produce direct knowledge to apply strategies in appropriate areas and thus is essential for N2O along the process. sustainable local food production. N2O emission maps, together with the The highest direct N2O emissions from cereal crops in the AMB came fertilizer use maps developed in this study, are already two key pieces of from Calcaric Cambisol I Haplic Calcisol soils. According to the Soil Atlas of the full N-Cycle for the city local food production, which should cover all Europe, Cambisols are deep, productive soils in Southern Europe and are N flows to air, water, soil and crops. Moreover, reducing N2O emissions primarily used to produce food/oil in Mediterranean regions (JRC, 2005). from peri-urban agriculture, by for instance managing crop type, fertiliza- For the AMB, over half of the cereals were produced in the Vallès Occiden- tion and irrigation in relation to soil type, also contribute to climate change tal region. Total cereals cultivation in the AMB covered 2024 ha and emit- mitigation. ted 3237 kg N2O-N year−1 (via Tier 1) and 1050 kg N2O-N year−1 (via Tier Expansion scenarios of peri-urban agricultural areas, as proposed by the 2). About 73% of the cereal areas in the AMB are rainfed and 27% are irri- PDU of the AMB (Barcelona Regional and AMB-PDU, 2019), show that re- gated, according to the URBAG map. Idescat (2015) meteorology report, storing agricultural areas lost in the past helps mitigate negative shows the Vallès Occidental region experienced an annual average temper- socioecological impacts by increasing diversity of ecosystem services e.g. ature of 15.5 °C and an annual mean rainfall of 326.6 mm. Emissions from food security and reducing reliance on imports (Padró et al., 2020). There- cereal crops were found to be much lower than those of horticultural crops, fore, appropriate management of new peri-urban agricultural areas could all irrigated. This could be related to the temperature, irrigation practice, bring along these advantages and generate synergies between peri-urban and fertilizer differences. Barton et al. (2008) showed that direct N2O emis- agriculture and environmental services such as climate change mitigation. sions are sensible to seasonal changes in temperature and rainfall. For ex- The way we calculate emissions and maps estimated here, may be used in ample, in a rainfed system there is not as much water flowing into the conjunction with other methods e.g. life cycle assessment and environmen- pores of agricultural soils, allowing space for available oxygen. Along tal risk assessment for broader sustainability assessments of for instance, with colder temperatures, these conditions are unsuitable for denitrifica- the PDU scenarios of expansion of peri-urban agriculture areas and the sus- tion. Furthermore, these conditions make nitrification the primary process tainability challenges brought by the expansion. for direct N2O production (Aguilera et al., 2013). 5. Conclusions 4.4. Implications for food sovereignty and sustainable peri-urban crop production This study presents improved direct N2O emissions maps from peri- N2O emissions maps of peri-urban agriculture activity, as we propose in urban agriculture in the Metropolitan Area of Barcelona (AMB). We recom- this study, are a key piece for an appropriate introduction of mitigation, mended using crop-specific EFs when calculating direct N2O emissions at 8 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 Fig. 5. Difference between emissions calculated with Tier 1 and Tier 2 emission factors, 100 × 100 meter resolution map. finer scales – such as those of peri-urban agriculture, because they cap- contributing to the AMB self-sufficiency and its resilience capacity, all ob- ture variability of crop management practices such as fertilization and jectives of the PDU. irrigation, relevant for the quantification of these emissions. Applying On-site data of crop management practices could help overcome uncer- crop-specific EFs and mapping emissions enables visualization of key tainties introduced due to data representativeness in this study. More de- emitting core areas and helps highlight appropriate strategies for their tailed climate dependant EFs, that reflects the soil and water influence on further mitigation. As a result, two core areas of emissions in the AMB N2O emissions generation, which are qualitatively accounted for in this were observed: the Baix Llobregat and Vallès Occidental regions. study, could help improve further the EFs used here. Mapping efforts that Overirrigation to crops may influence emissions especially in the Baix incorporate local data of water management and soil characteristics can Llobregat region, and over fertilization also influence emissions in also allow for indirect N2O emissions to be estimated. Furthermore, this both of these areas where irrigated vegetables and rainfed cereals dom- case study provides rough estimations of N2O emissions for the AMB and inate, according to our improved agricultural land use map. Results should be seen as a point of reference i.e. a first picture, to further build show that Tier 1 leads to lower emissions for irrigated crops compared upon more robust emission inventories and improved EFs also accounting to crop-specific EFs, while it leads to higher emissions for rainfed for temporal variability of climate and crop management practices. crops, highlighting the need to account for local parameters while quan- tifying N2O emissions in regional studies. CRediT authorship contribution statement This study's data and method to map and estimate direct N2O emissions can be further applied to areas beyond peri-urban agriculture areas as long Conceptualization: AMB, KJ, GV. Data collection: AMB, KJ, SV, CML. as the cultivation techniques are similar to those of conventional agricul- Data analysis: AMB, KJ, MRS. Visualizations: AMB, KJ. Writing, review ture, as defined here i.e. open-air, soil-based, ground-base, conventional ag- and editing of manuscript: AMB, KJ, MRS, SV, CML, GV. Supervision: GV. ricultural practices. For the case of organic, hydroponic or other hi-tech urban agriculture, emission factors and fertilizer rates do not hold, and Declaration of competing interest should be updated despite that similar mapping and estimation of emis- sions could be used. The authors declare that they have no known competing financial inter- Ideally, spatially distributing all nitrogen flows i.e. all inputs (including ests or personal relationships that could have appeared to influence the organic fertilization) and outputs (including all N-related emissions), could work reported in this paper. help identify new opportunities to increase the circularity of the sector, apply appropriate strategies towards sustainability and mitigate climate Acknowledgments change. For instance, optimizing the use of fertilizers and irrigation for the peri-urban agriculture in the AMB, can actively contribute to the mitiga- This work has been made possible thanks to the financial support of the tion of N2O emissions and thus contribute as a strategy for climate change ERC Consolidator Integrated System Analysis of Urban Vegetation and Ag- mitigation and help tackle the current climate crisis, besides the importance riculture (818002-URBAG) and the funding from the research and innova- of the local food supplied to the city from these areas that are actively tion programme under the H2020 Marie Skłodowska-Curie Actions 9 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 PROTEAN project (842460). This work contributes to the “María de Gerber, J.S., Carlson, K.M., Makowski, D., Mueller, N.D., Garcia de Cortazar-Atauri, I., Havlík, P., Herrero, M., Launay, M., O’Connell, C.S., Smith, P., West, P.C., 2016. Spatially explicit Maeztu” Programme for Units of Excellence of the Spanish Ministry of estimates of N2O emissions from croplands suggest climate mitigation opportunities from Science and Innovation (CEX2019-000940-M) at ICTA-UAB. improved fertilizer management. Glob. Chang. Biol. 22, 3383–3394. https://doi.org/10. 1111/gcb.13341. Gold, C., Angel, P., 2006. Voronoi hierarchies. Lecture Notes in Computer Science (Including Appendix A. Supplementary data Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 99–111 https://doi.org/10.1007/11863939_7. Supplementary data to this article can be found online at https://doi. Guevara, M., Tena, C., Porquet, M., Jorba, O., Pérez García-Pando, C., 2019. HERMESv3, a stand-alone multi-scale atmospheric emission modelling framework – part 1: global and org/10.1016/j.scitotenv.2022.153514. regional module. Geosci. Model Dev. 12, 1885–1907. https://doi.org/10.5194/gmd-12- 1885-2019. ICGC, 2018. Mapa de sòls 1:250.000. URL https://www.icgc.cat/Administracio-i-empresa/D escarregues/Cartografia-geologica-i-geotematica/Cartografia-de-sols/Mapa-de-sols-1- References 250.000. (Accessed 11 May 2021) (WWW Document). Idescat, 2015. Anuario estadístico de Cataluña. URL https://www.idescat.cat/pub/? Aguilera, E., Lassaletta, L., Sanz-Cobena, A., Garnier, J., Vallejo, A., 2013. The potential of or- id=aec&n=445&t=2016&lang=es. (Accessed 11 May 2021) (WWW Document). ganic fertilizers and water management to reduce N2O emissions in Mediterranean cli- Institut Cartogràfic i Geològic de Catalunya, 2020. Base Municipal. URL https://www.icgc. mate cropping systems. A review. Agric. Ecosyst. Environ. 164, 32–52. https://doi.org/ cat/Administracio-i-empresa/Descarregues/Capes-de-geoinformacio/Base-municipal 10.1016/j.agee.2012.09.006. (WWW Document). Albanito, F., Lebender, U., Cornulier, T., Sapkota, T.B., Brentrup, F., Stirling, C., Hillier, J., IPCC, Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, 2017. Direct nitrous oxide emissions from tropical and sub-tropical agricultural systems M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, - a review and modelling of emission factors. Sci. Rep. 7, 44235. https://doi.org/10. T., Yelekçi, O., Zhou, R., Y.B., 2021. In: MassonDelmotte, V., Zhai, P. (Eds.), Climate 1038/srep44235. Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Aljoumani, B., 2012. Soil Water Management: Evaluation of Infiltration in Furrow Irrigation Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Uni- Systems, Assessing Water and Salt Content Spatially and Temporally in the Parc Agrari versity Press (IPCC). del Baix Llobregat Area. Universitat Politècnica de Catalunya (TDX (Tesis Doctorals en JRC, 2005. Soil Atlas of Europe. URL https://publications.jrc.ec.europa.eu/repository/ Xarxa)). handle/JRC32396. (Accessed 11 May 2021) (WWW Document). Arcas-Pilz, V., Parada, F., Villalba, G., Rufí-Salis, M., Rosell-Melé, A., Gabarrell Durany, X., Langemeyer, J., Madrid-Lopez, C., Mendoza Beltran, A., Villalba Mendez, G., 2021. 2021. Improving the fertigation of soilless urban vertical agriculture through the combi- Urban agriculture — a necessary pathway towards urban resilience and global nation of struvite and rhizobia inoculation in Phaseolus vulgaris. Front. Plant Sci. 12, 952. sustainability? Landsc. Urban Plan. 210, 104055. https://doi.org/10.1016/j. https://doi.org/10.3389/FPLS.2021.649304/BIBTEX. landurbplan.2021.104055. Barcelona Regional, AMB-PDU, 2019. Initial Strategic Document. Advance of the Urbanisitic Lassaletta, L., Billen, G., Grizzetti, B., Anglade, J., Garnier, J., 2014. 50 year trends in nitrogen Metropolitan Master Plan (Barcelona, Spain). use efficiency of world cropping systems: the relationship between yield and nitrogen Barton, L., Kiese, R., Gatter, D., Butterbach-bahl, K., Buck, R., Hinz, C., Murphy, D.V., 2008. input to cropland. Environ. Res. Lett. 9, 105011. https://doi.org/10.1088/1748-9326/ Nitrous oxide emissions from a cropped soil in a semi-arid climate. Glob. Chang. Biol. 9/10/105011. 14, 177–192. https://doi.org/10.1111/j.1365-2486.2007.01474.x. Leip, A., Busto, M., Winiwarter, W., 2011. Developing spatially stratified N 2O emission fac- Bouwman, A.F., Boumans, L.J.M., Batjes, N.H., 2002. Emissions of N2O and NO from fertil- tors for Europe. Environ. Pollut. 159, 3223–3232. https://doi.org/10.1016/j.envpol. ized fields: summary of available measurement data. Glob. Biogeochem. Cycles 16, 2010.11.024. 1058. https://doi.org/10.1029/2001gb001811. MAPA, 2015. Surfaces and annual production of crops. URL https://www.mapa.gob.es/es/ Brentrup, F., Kusters, J., Lammel, J., Kuhlmann, H., 2000. Methods to estimate on-field nitro- estadistica/temas/estadisticas-agrarias/agricultura/superficies-producciones-anuales- gen emissions from crop production as an input to LCA studies in the agricultural sector. cultivos/. (Accessed 13 May 2021) (WWW Document). Int. J. Life Cycle Assess. 5, 349–357. https://doi.org/10.1007/bf02978670. MAPA, 2016. Balance del Nitrogeno en la Agricultural Espanola. URL https://www.mapa. Cayuela, M.L., Aguilera, E., Sanz-Cobena, A., Adams, D.C., Abalos, D., Barton, L., Ryals, gob.es/es/agricultura/temas/medios-de-produccion/bn2016_metodologia- R., Silver, W.L., Alfaro, M.A., Pappa, V.A., Smith, P., Garnier, J., Billen, G., resultados_tcm30-507806.pdf (WWW Document). Bouwman, L., Bondeau, A., Lassaletta, L., 2017. Direct nitrous oxide emissions in Mercade, L., Delgado, M. del M., Gil, J.M., 2010. Pràctiques de Fertilització a Catalunya. Mediterranean climate cropping systems: emission factors based on a meta- Enquesta 2010. analysis of available measurement data. Agric. Ecosyst. Environ. 238, 25–35. Metropolitan Area of Barcelona, 2021. Pla Director Urbanistic Metropolita. URL https:// https://doi.org/10.1016/j.agee.2016.10.006. amb1.maps.arcgis.com/apps/Cascade/index.html? Corre, M.D., Schnabel, R.R., Shaffer, J.A., 1999. Evaluation of soil organic carbon under for- appid=4ea47987014c45b586b4b9c4a7467adb. (Accessed 13 May 2021) (WWW Docu- ests, cool-season and warm-season grasses in the northeastern US. Soil Biol. Biochem. ment). 31, 1531–1539. https://doi.org/10.1016/S0038-0717(99)00074-7. Monfreda, C., Ramankutty, N., Foley, J.A., 2008. Farming the planet: 2. Geographic distribu- CREAF, 2015. Land cover map of Catalonia. URL http://www.creaf.uab.es/mcsc/usa/index. tion of crop areas, yields, physiological types, and net primary production in the year htm (WWW Document). 2000. Glob. Biogeochem. Cycles 22, 1022. https://doi.org/10.1029/2007GB002947. Dachraoui, M., Sombrero, A., 2020. Effect of tillage systems and different rates of nitrogen Mueller, N.D., Gerber, J.S., Johnston, M., Ray, D.K., Ramankutty, N., Foley, J.A., 2012. Clos- fertilisation on the carbon footprint of irrigated maize in a semiarid area of Castile and ing yield gaps through nutrient and water management. Nature 490, 254–257. https:// Leon, Spain. Soil Tillage Res. 196, 104472. https://doi.org/10.1016/j.still.2019.104472. doi.org/10.1038/nature11420. DARPA, 2015. Mapa de cultius DUN-SIGPAC. URL http://agricultura.gencat.cat/ca/serveis/ Muller, A., Schader, C., El-Hage Scialabba, N., Brüggemann, J., Isensee, A., Erb, K.H., Smith, cartografia-sig/aplicatius-tematics-geoinformacio/sigpac/mapa-cultius/. (Accessed 11 P., Klocke, P., Leiber, F., Stolze, M., Niggli, U., 2017. Strategies for feeding the world May 2021) (WWW Document). more sustainably with organic agriculture. Nat. Commun. 8, 1–13. https://doi.org/10. De Klein, C., Novoa, R.S.A., Ogle, S., Smith, K.A., Rochette, P., Wirth, T.C., McConkey, B.G., 1038/s41467-017-01410-w. Mosier, A., Rypdal, K., Walsh, M., Williams, S.A., 2006. Chapter 11: N2O emissions Padró, R., La Rota-Aguilera, M.J., Giocoli, A., Cirera, J., Coll, F., Pons, M., Pino, J., Pili, S., Ser- from managed Soils, and CO2 emissions from lime and urea application. IPCC Guidelines rano, T., Villalba, G., Marull, J., 2020. Assessing the sustainability of contrasting land use for National GHG Inventories. scenarios through the Socioecological Integrated Analysis (SIA) of the metropolitan green Dirección General del catastro, 2014. Modelo de datos de cartografía vectorial (formato infrastructure in Barcelona. Landsc. Urban Plan. 203, 103905. https://doi.org/10.1016/j. shapefile). URL http://www.catastro.meh.es/ayuda/manual_descriptivo_shapefile.pdf landurbplan.2020.103905. (WWW Document). Peñuelas, J., Poulter, B., Sardans, J., Ciais, P., Van Der Velde, M., Bopp, L., Boucher, O., Dobbie, K.E., McTaggart, I.P., Smith, K.A., 1999. Nitrous oxide emissions from intensive agri- Godderis, Y., Hinsinger, P., Llusia, J., Nardin, E., Vicca, S., Obersteiner, M., Janssens, cultural systems: variations between crops and seasons, key driving variables, and mean I.A., 2013. Human-induced nitrogen-phosphorus imbalances alter natural and managed emission factors. J. Geophys. Res. Atmos. 104, 26891–26899. https://doi.org/10.1029/ ecosystems across the globe. Nat. Commun. 4, 1–10. https://doi.org/10.1038/ 1999JD900378. ncomms3934. Dorr, E., Goldstein, B., Horvath, A., Aubry, C., Gabrielle, B., 2021. Environmental impacts and Pi, Deng, J., 2019. Final Report Improving DNDC Modeling Capability to Quantify Mitigation resource use of urban agriculture: a systematic review and meta-analysis. Environ. Res. Potential of Nitrous Oxide From California Agricultural Soils. Lett. 16, 093002. https://doi.org/10.1088/1748-9326/ac1a39. Postigo, C., Ginebreda, A., Barbieri, M.V., Barceló, D., Martín-Alonso, J., de la Cal, A., Boleda, Flynn, H.C., Smith, J., Smith, K.A., Wright, J., Smith, P., Massheder, J., 2005. Climate- and M.R., Otero, N., Carrey, R., Solà, V., Queralt, E., Isla, E., Casanovas, A., Frances, G., López crop-responsive emission factors significantly alter estimates of current and future nitrous de Alda, M., 2021. Investigative monitoring of pesticide and nitrogen pollution sources in oxide emissions from fertilizer use. Glob. Chang. Biol. 11, 1522–1536. https://doi.org/ a complex multi-stressed catchment: the lower Llobregat River basin case study (Barce- 10.1111/j.1365-2486.2005.00998.x. lona, Spain). Sci. Total Environ. 755, 142377. https://doi.org/10.1016/j.scitotenv. García-Serrano Jiménez, P., Lucena Marotta, J.J., Ruano Criado, S., Nogales García, M., López 2020.142377. Bellido, L., Betrán Aso, J., Ramos Monreal, Á., López Córcoles, H., López Fuster, P., Rochette, P., Worth, D.E., Lemke, R.L., McConkey, B.G., Pennock, D.J., Wagner-Riddle, C., Bermejo Corrales, J.L., Urbano Terrón, P., Piñeiro Andión, J., Castro Insua, J., Blázquez Desjardins, R.L., 2008. Estimation of N2O emissions from agricultural soils in Canada. Rodríguez, R., Ramos Mompó, C., Pomares García, F., Quiñones Oliver, A., Martínez I. Development of a country-specific methodology. Can. J. Soil Sci. 88, 641–654. Alcántara, B., Primo-Millo, E., Legaz Paredes, F., Espada Carbó, J.L., García-Escudero https://doi.org/10.4141/CJSS07025. Domínguez, E., García García, C., Rodríguez Pérez, J., 2010. Guia Practica de la Rufí-Salís, M., Garcia-Orellana, J., Cantero, G., Castillo, J., Hierro, A., Rieradevall, J., Bach, J., Fertilizacion Racional de los Cultivos en España. Part I y II. 2019. Influence of land use changes on submarine groundwater discharge. Environ. Res. Generalitat de Catalunya, 2016. Superfícies, rendiments i produccions del conreus agrícoles. Commun. 1, 031005. https://doi.org/10.1088/2515-7620/ab1695. 10 A. Mendoza Beltran et al. Science of the Total Environment 822 (2022) 153514 Rufí-Salís, M., Brunnhofer, N., Petit-Boix, A., Gabarrell, X., Guisasola, A., Villalba, G., 2020. Vittuari, M., Bazzocchi, G., Blasioli, S., Cirone, F., Maggio, A., Orsini, F., Penca, J., Petruzzelli, Can wastewater feed cities? Determining the feasibility and environmental burdens of M., Specht, K., Amghar, S., Atanasov, A.-M., Bastia, T., Bertocchi, I., Coudard, A., struvite recovery and reuse for urban regions. Sci. Total Environ. 737, 139783. https:// Crepaldi, A., Curtis, A., Fox-Kämper, R., Gheorghica, A.E., Lelièvre, A., Muñoz, P., doi.org/10.1016/j.scitotenv.2020.139783. Nolde, E., Pascual-Fernández, J., Pennisi, G., Pölling, B., Reynaud-Desmet, L., Righini, Serra, P., Salvati, L., Queralt, E., Pin, C., Gonzalez, O., Pons, X., 2016. Estimating water con- I., Rouphael, Y., Saint-Ges, V., Samoggia, A., Shaystej, S., da Silva, M., Toboso Chavero, sumption and irrigation requirements in a long-established Mediterranean rural commu- S., Tonini, P., Trušnovec, G., Vidmar, B.L., Villalba, G., De Menna, F., 2021. Envisioning nity by remote sensing and field data. Irrig. Drain. 65, 578–588. https://doi.org/10. the future of European food systems: approaches and research priorities after COVID- 1002/ird.1978. 19. Front. Sustain. Food Syst. 5, 1–9. https://doi.org/10.3389/fsufs.2021.642787. Serra, P., Saurí, D., Salvati, L., 2018. Peri-urban agriculture in Barcelona: outlining landscape Wang, J., Xiong, Z., Yan, X., 2011. Fertilizer-induced emission factors and background emis- dynamics vis à vis socio-environmental functions. Landsc. Res. 43, 613–631. https://doi. sions of N2O from vegetable fields in China. Atmos. Environ. 45, 6923–6929. https://doi. org/10.1080/01426397.2017.1336758. org/10.1016/j.atmosenv.2011.09.045. Shepherd, A., Yan, X., Nayak, D., Newbold, J., Moran, D., Dhanoa, M.S., Goulding, K., Smith, Wang, Q., Zhou, F., Shang, Z., Ciais, P., Winiwarter, W., Jackson, R.B., Tubiello, F.N., P., Cardenas, L.M., 2015. Disaggregated N2O emission factors in China based on cropping Janssens-Maenhout, G., Tian, H., Cui, X., Canadell, J.G., Piao, S., Tao, S., 2020. Data- parameters create a robust approach to the IPCC Tier 2 methodology. Atmos. Environ. driven estimates of global nitrous oxide emissions from croplands. Natl. Sci. Rev. 7, 122, 272–281. https://doi.org/10.1016/j.atmosenv.2015.09.054. 441–452. https://doi.org/10.1093/nsr/nwz087. Tesfaye, K., Takele, R., Sapkota, T.B., Khatri-Chhetri, A., Solomon, D., Stirling, C., Albanito, F., Yue, Q., Wu, H., Sun, J., Cheng, K., Smith, P., Hillier, J., Xu, X., Pan, G., 2019. Deriving emis- 2021. Global high-resolution gridded dataset of N2O emission and mitigation potential sion factors and estimating direct nitrous oxide emissions for crop cultivation in China. from maize and wheat fields. Data Brief 37, 107239. https://doi.org/10.1016/J.DIB. Environ. Sci. Technol. 53, 10246–10257. https://doi.org/10.1021/acs.est.9b01285. 2021.107239. UNFCCC, 1997. Kyoto Protocol (Kyoto). UNFCCC, 2015. Paris Agreement, Conference of the Parties on Its Twenty-first Session (doi: FCCC/CP/2015/L.9/Rev.1). 11