WEP – Wine Economics and Policy Just Accepted Manuscript 1 Just accepted 1 2 3 4 How are vineyards management strategies and climate-related conditions 5 affecting economic performance? A case study of Chilean wine grape growers 6 7 Carlos Bopp1,3, Roberto Jara-Rojas1,3, Alejandra Engler2,3, Miguel Araya-Alman4 8 9 10 11 12 1 Department of Agricultural Economics, Universidad de Talca, Chile. Postal code: 3460000 13 2 Department of Agricultural Economics, Pontificia Universidad Católica de Chile, Santiago, Chile. 14 Postal code: 7820436 15 3 Researcher Núcleo Milenio CESIEP, Chile. Postal code: 7820436 16 4 Department of Agricultural Sciences, Universidad Católica del Maule, Chile. Postal code: 3340000 17 18 Correspondence concerning this article should be addressed to Roberto Jara-Rojas, Department of 19 Agricultural Economics, Universidad de Talca, Chile. Postal code: 3460000, E-mail: rjara@utalca.cl; 20 Tel.: +56 71 2200214. 21 22 This article has been accepted for publication and undergone full peer review but has not been through 23 the copyediting, typesetting, pagination and proofreading process, which may lead to differences 24 between this version and the Version of Record. 25 26 Please cite this article as: 27 28 Bopp C., Jara-Rojas R., Engler A., Araya-Alman M. (2022), How are vineyards management 29 strategies and climate-related conditions affecting economic performance? A case study of Chilean 30 wine grape growers, Wine Economics and Policy, Just Accepted. 31 DOI: 10.36253/wep-12739 32 33 34 35 36 mailto:rjara@utalca.cl WEP – Wine Economics and Policy Just Accepted Manuscript 2 Abstract 37 In wine grape production, growers decide between alternative management strategies of the vineyard 38 that have direct consequences on competitiveness. The aim of this study is to evaluate the impact on 39 economic performance of four management strategies: training system, reserve quality production, 40 irrigation method, and mechanization of labors. The data used in the study comes from face-to-face 41 interviews to 336 wine grape growers of Central Chile, which was complemented with climatic 42 variables retrieved from Geographic Information Systems. A log-log regression model of total value 43 product (TVP) for the main variety grown in the vineyard was estimated, using production factors, 44 vineyards’ attributes, management strategies and climate-related conditions as explanatory variables. 45 An interesting contribution of this study is the identification of TVP functions for land, fertilizers, 46 fungicides, other agrochemicals, labor, and age of vines. Our results show that the training system 47 has the most impact on TVP, where tendone-trained vineyards demonstrated 50% higher TVP than 48 those vertically trained. Reserve quality production also has a positive effect on TVP, increasing it 49 by 22% compared to vineyards producing varietal quality grapes. In contrast, the use of pressurized 50 irrigation systems and mechanization in harvesting do not present a significant effect on TVP. The 51 findings of this paper represent an advance in the understanding of the economic performance factors 52 associated with wine grape growing and could serve to guide on-farm decisions and sectoral policies 53 in pursuing the competitive development of wine grape growers. 54 55 Keywords: Economic performance, production function, vineyard management, wine grape growing 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 WEP – Wine Economics and Policy Just Accepted Manuscript 3 1. Introduction 71 One of the main components of competitiveness in wine grape production lies in the capacity to 72 innovate [1] and to improve performance using available resources [2, 3]. The process of innovation 73 at the vineyard level has played a prominent role in emerging countries from South America, South 74 Africa, Asia and Oceania [4, 5, 6]. These countries have expanded their vineyard production, albeit 75 not neglecting wine quality, to the extent that they are not only challenging the old world’s leaders 76 but also are increasing their domestic market share [7, 8, 9, 10]. Hence, there is evidence of 77 improvements in competitiveness because of technological modernization processes, which has been 78 especially relevant in developing countries. 79 An interesting example of this is Chile, a South American country that has experienced rapid 80 development of its export-oriented wine industry in recent decades [11]. Indeed, wine grapes are one 81 of the most important crops in the country [12]. Between 1990 and 2015, vineyard plantations 82 doubled, wine production increased fivefold, and wine export volume grew from 22 to 1,445 million 83 liters [13]. As a result, Chile has become an important player in international markets, being an 84 example of how a traditional industry can become highly competitive in a short period of time by 85 implementing important changes in technologies and production systems. 86 Despite the overall progress of the Chilean wine grape industry, there are some concerns in the 87 domestic market from producers’ associations regarding an oligopsony market structure (i.e., few 88 grape buyers) that would generate competitiveness problems for wine grape growers [14]. For that 89 reason, on-farm competitiveness has turned to be an extremely relevant issue for the viticultural sector 90 and a better understanding is required of the factors affecting vineyards’ economic performance, such 91 as the impact of innovations and management strategies. In this regard, management strategies are 92 considered among the most important determinants of vineyard profitability [3, 15, 16, 17]. Within 93 this category we distinguish between production technologies, such as pressurized irrigation or 94 mechanization in harvesting, that are generally more affordable for larger producers because of 95 economies of scale and financial access [3], and cultivation techniques, such as training systems and 96 reserve quality growing, that are generally less demanding in financial capital. 97 This study seeks to understand the role of vineyards management strategies on the economic outcome 98 exhibited by wine grape growers, controlling for other production factors (e.g., land, labor, and 99 inputs) and climate-related conditions (i.e., potential evapotranspiration, precipitation, and chilling 100 hours). Using Chile as a case study, the aim of this paper is to provide insights about vineyard-level 101 drivers of competitive performance in emerging countries. Prior research analysing vineyards 102 outcomes related to economic performance, efficiency, or productivity, have focused mainly on the 103 effect of economies of scale [5, 10, 18]; to the best of our knowledge, there are no studies analyzing 104 WEP – Wine Economics and Policy Just Accepted Manuscript 4 management strategies implemented by wine grape growers in explaining economic performance. 105 The study of Urso et al. [19] is one of the few that evaluates several production unit and contextual 106 factors of vineyards; however, it is focused on production efficiency rather than analyzing the 107 contribution of growers’ production decisions on performance. Instead, our paper examines to what 108 extent management strategies implemented by wine grape growers affect the TVP at the vineyard 109 level, considering the heterogeneity of production units’ attributes and climate-related conditions 110 under which they operate. 111 The vineyards management strategies analyzed in this study were: a) training system (tendone vs. 112 vertical structures), b) wine grape destination (reserve vs. varietal wines), c) irrigation method 113 (pressurized vs. gravity irrigation), and d) mechanization in harvesting (mechanized vs. hand-picked). 114 These vineyards’ strategies are of different scope and nature, some of them represent structural (fixed) 115 decisions while others are more related to flexible (alternative) decisions. For instance, wine grape 116 destination is a flexible decision that might be defined each season, though it involves an array of 117 practices aiming to regulate vine yield and grape quality, such as canopy management (e.g., 118 pruning/mooring, de-sprouting, canopy defoliation, tipping of shoots) [20, 21], agrochemical use and 119 irrigation regimes, among others. In contrast, the training system is a structural decision that must be 120 made when wine grape growers establish the vineyard and is not (easily) modifiable. 121 The paper is structured as follows. The next section details the data used to perform the analysis and 122 finishes with the empirical model. The third section presents and discusses results, and the last section 123 summarizes the most relevant conclusions of the study. 124 125 2. Materials and Methods 126 127 2.1. Sampling procedure and data collection 128 The study area covers the O’Higgins and Maule regions in Central-South Chile (33° 50’ and 36° 33’ 129 S, WGS84 datum), located in central Chile in the heart of the fruit and vineyard production (Figure 130 1). Combined, both regions comprise 73% of the national planted area of vineyards, distributed 131 among three important valleys, from north to south: Rapel, Curicó, and Maule (a brief description of 132 the weather conditions prevailing in these valleys is presented in Appendix 1). The area under study 133 has a temperate Mediterranean climate, characterized by a six month dry season (Sept- Mar) and a 134 rainy winter, with precipitation between 600 and 700 mm annually. The primary data used in this 135 study was generated at the vineyard level, administering a georeferenced survey on-site to 436 wine 136 grape growers between October 2014 and March 2015. This survey was restricted to vineyards from 137 irrigated lands, growing at least one hectare. The sampling procedure consisted of a stratified random 138 WEP – Wine Economics and Policy Just Accepted Manuscript 5 sample across 16 municipalities, where the number of surveys administered was determined 139 depending on the relative number of vineyards in each municipality. The municipalities were, in order 140 of number of surveyed producers: San Javier, Sagrada Familia, Curicó, Nancagua, Villa Alegre, Santa 141 Cruz, Talca, Palmilla, San Clemente, Peralillo, Río Claro, Requínoa, Chimbarongo, Maule, San 142 Vicente, and Peumo. After the field data collection process, in September 2020, using the 143 georeferenced point of each survey, the dataset was supplemented with spatialized data of climate-144 related conditions 2015/2016 from the Chilean Natural Resources Information Center (CIREN) [22]. 145 CIREN is a public institution that provides information on the natural and productive resources of the 146 country through the use of geospatial data and applications. In this paper, the data from CIREN 147 referred uniquely to environmental information for the years 2015-2016. As result of merging the 148 primary and secondary data, the final sample with complete information was reduced to 336 149 observations because the Geographic Information System (GIS) used in this study did not cover the 150 total distribution of surveyed vineyards. 151 152 153 Figure 1. Map of the study area and locations of the vineyards included in the sample (black dots). 154 155 2.2. Survey data 156 The questionnaire administered to wine grape growers collected detailed economic and agronomic 157 information for the main variety grown in the vineyard, such as planted area, yield, grape price, and 158 WEP – Wine Economics and Policy Just Accepted Manuscript 6 (per hectare) intensity of use of inputs and labor. Growers were asked about the number of 159 applications, doses, and unitary prices in the case of agrochemicals (i.e., fertilizers, herbicides, 160 insecticides, fungicides, and acaricides) and number of working-days or agricultural 161 machines/equipment in the case of labor (i.e., harvest, pruning/mooring, tipping of shoots, de-162 sprouting, canopy defoliation, physical weed control, and other labor), which were valued at fixed 163 market prices. 164 Regarding growers’ performance, the yield obtained by each grower (kg ha-1) was multiplied by the 165 average grape price of the variety in the sample ($ kg-1). As in our sample growers identified 19 166 different varieties, we used the average price for each variety to estimate their incomes. The reason 167 for using fixed grape prices and fixed market prices for inputs and labor was to avoid differences in 168 bargaining power or personal skills among wine grape growers, which are beyond the scope of our 169 analysis as the objective of our paper is to estimate the impact of technical decisions on technical 170 outcomes using an economic model. 171 Subsequently, to convert the monetary measures per hectare for inputs, labors, and output to the plot 172 level, they were scaled-up (values were multiplied by the planted area of the main variety grown in 173 the vineyard). Hence, the economic output variable analyzed in this paper is the total value product 174 (TVP) generated by the main variety of the vineyard, considering that there are important differences 175 in prices between grape varieties within the sample. For the purposes of this study, expenditures and 176 total value products were converted to US dollars using the average exchange rate of 2015 (654 177 Chilean pesos per US dollar), the year in which the field survey process finished. 178 179 2.3. GIS spatial data 180 An important feature of this study is the inclusion of climate-related variables as controls in the 181 econometric model. In particular, we included three variables: potential evapotranspiration, 182 precipitation, and chilling hours; a description is presented in Table 1. The selection of these 183 variables, representing referential production conditions for vineyards, is expected to exert an 184 influence on vineyard yields. The climate-related variables were retrieved from high spatial resolution 185 data of the O’Higgins and Maule regions of Chile, using layers and isolines of Agroclimatic Districts 186 (1:250,000 scale) gathered from the Chilean Natural Resources Information Center (CIREN) [22]. 187 An intersection algorithm able to cross climatic layers and the georeferenced sampling site of each 188 vineyard allowed us to add secondary information to our dataset of surveyed wine grape growers. 189 This procedure was performed using the QGIS software (Open-Source Geospatial Foundation 190 Project: http://qgis.osgeo.org). 191 192 http://qgis.osgeo.org/ WEP – Wine Economics and Policy Just Accepted Manuscript 7 3. Calculation 193 According to Chinnici et al. [23], evaluating the operational choices of a vineyard involves knowledge 194 of the potentials and restrictions of both a technical and economic-managerial nature. Indeed, growers 195 face different alternatives in which to invest but they have certain restrictions imposed by their own 196 attributes and other territorial characteristics, ranging from natural resources to the availability of 197 production factors and techniques [1]. Therefore, this paper considers that growers’ TVP is a function 198 of production factors (i.e., land, input, labor) attributes of the productive unit, climate-related 199 variables, and management strategies. 200 To model the TVP generated by wine grape growers, we adopted a Cobb-Douglas functional form 201 estimated using a multiple linear regression, in logarithms for all continuous variables. The empirical 202 model in natural logarithms for the i-th wine grape grower can be expressed as follows: 203 𝑙𝑛 𝑌𝑖 = 𝛼 + ∑ 𝛽𝑗𝑖 𝑙𝑛 𝑋𝑗𝑖 5 𝑗=1 + ∑ 𝛾𝑘𝑖 𝐴𝑘𝑖 3 𝑘=1 + ∑ 𝜑𝑚𝑖 𝑀𝑚𝑖 4 𝑚=1 + ∑ 𝜋𝑙𝑖 𝐸𝑙𝑖 3 𝑙=1 + 𝑣𝑖 (Eq. 1) 204 The dependent variable in our study is the total value product of wine grape growers (Y), which 205 comes from the multiplication of yields (kg ha-1) per planted area (ha) and grape price ($ kg-1). The 206 model is expressed as a function of five inputs: Land (X1), Fertilizers (X2), Fungicides (X3), Other 207 agrochemicals (X4), and Labor expenditures (X5). In the case of other agrochemicals, this category 208 represents the sum of expenditures in insecticides, acaricides, and herbicides; fertilizers and 209 fungicides were incorporated in isolation into the model because of their agronomic importance in 210 vineyard production. In the empirical model, there are also three sets of control variables for: a) 211 attributes of the productive unit, b) climate-related variables, and c) management strategies. First, a 212 set of three variables representing productive unit attributes was considered: grape color (A1), age of 213 the vines (A2), and valley where the vineyard is located (A3). Following, a set of four dummy 214 variables for management strategies: pressurized irrigation (M1) and mechanized harvest (M2), 215 training system (M3), and type of wine for which the grapes are intended (M4). And finally, a set of 216 three climate-related variables, namely: Potential evapotranspiration (E1), Precipitation (E2), and 217 Chilling hours (E3). The last term of equation 1, v_i, is the normally distributed error that accounts 218 for statistical noise in the model. 219 To test the robustness of our empirical model and observe the contribution of the different sets of 220 variables included in the model, several progressive specifications for the above explained sets of 221 explanatory variables were estimated and compared through maximum likelihood ratio tests. A 222 complete explanation of the covariates included in the equations is shown in Table 1. The described 223 model was estimated in STATA 15.1 [24]. 224 WEP – Wine Economics and Policy Just Accepted Manuscript 8 225 4. Results and Discussion 226 227 4.1. Vineyards’ total value product and explanatory variables 228 Table 1 presents a description and summary statistics of the variables included in the models. It is 229 worth noting that values are reported for the main grape variety at the plot level. 230 231 Table 1. Variable description and summary statistics of variables used in models of vineyard 232 production for three wine grape growing areas of Chile (data at the plot level for the main grape 233 variety of the vineyard; N= 336). 234 Variable Description Mea n S.D. Media n Mi n Max D V TVP Total value product (1,000 USD) 65.6 0 104.4 7 29.36 0.6 0 1213.7 6 P ro d u c ti o n f a c to rs Land Planted area (hectares) 16.7 4 20.28 9.90 1.0 0 140.00 Fertilizers Fertilizer expenditure (1,000 USD) 4.34 7.36 1.70 0.0 0 52.95 Fungicides Fungicide expenditure (1,000 USD) 2.89 5.63 0.99 0.0 0 51.38 Agrochem . Expenditure in agrochemicals to control insects, spiders and weeds (1,000 USD) 5.99 17.29 1.52 0.0 0 201.38 Labor Labor expenditure (1,000 USD) 16.4 9 21.05 8.13 0.2 8 137.61 V in e y a rd s’ a tt ri b u te s Grape Color Grape color (red=1; white=0) 0.82 0.38 1 0 1 Vineyard age Age of planting (years) 29.8 4 26.28 19 4 116 Rapel valley Rapel valley (yes=1; no= 0). 0.35 0.48 0 0 1 WEP – Wine Economics and Policy Just Accepted Manuscript 9 Curicó valley Curicó valley (yes=1; no= 0, excluded category in models) 0.20 0.40 0 0 1 Maule valley Maule valley (yes=1; no= 0). 0.45 0.50 0 0 1 M a n a g e m e n t st ra te g ie s Irrig. method Irrigation method (pressurized= 1; gravity= 0) 0.39 0.49 0 0 1 Mech. harv. Machinery use for harvest (yes= 1; no= 0) 0.17 0.38 0 0 1 Training syst. Training system (tendone=1; vertical=0) 0.18 0.39 0 0 1 Grape Dest Grape destination (reserve=1; varietal=0) 0.11 0.32 0 0 1 C li m a ti c c o n d it io n s Evapotran sp. Cumulative evapotranspiration from Dec-15 to Feb-16 (mm) 456 21 461 40 8 512 Precipitati on Cumulative precipitation from Dec-15 to Feb-16 (mm) 22.8 1 7.23 24 8 45 Chilling hours Cumulative chilling hours in 2016 (hours) 1,28 7 303 1,380 75 0 1,830 235 As shown in Table 1, growers’ TVP and input and labor expenditures exhibit considerable differences 236 between the mean and median, which reveals the skewed distribution to the left of these variables. 237 Planted area is also a skewed variable, where the mean surface is 16.7 ha, and the median is 9.9 ha. 238 The use of logarithms, besides its convenience in estimating partial elasticities of productive factors, 239 helps to avoid the skewed distribution of the data. 240 Turning to descriptive statistics, at median values at the plot level wine grape growers spent about 241 US$ 1,700, US$ 990 and US$ 1,520 on fertilizers, fungicides, and other agrochemicals, respectively. 242 The expenditure in labors – including harvest, pruning/mooring, tipping of shoots, de-sprouting, 243 canopy defoliation, physical weed control, and rest of labors – reached a median of US$ 8,130 in the 244 sample. The sum of expenditures on fertilizers, fungicides, other agrochemicals (to control insects, 245 spiders, and weeds), and labor represents an approximation of the operational costs incurred by grape 246 growers in a year, which reach a median value of US$15.005. On the other hand, the median TVP 247 was US$ 29,360. Note that the median planted area was 9.9 ha, which informs about an approximate 248 WEP – Wine Economics and Policy Just Accepted Manuscript 10 per hectare outcome of US$ 2,965 (this calculation is close to the actual median of the sample used 249 to estimate the model, which corresponds to USD$ 3,058 per hectare). 250 Regarding vineyards’ attributes, most wine grape growers cultivate red grapes (82%) rather than 251 white grapes (the remaining 18%). The median age of the vineyards was 19 years, within a range of 252 4 and 116 years old. Regarding wine valleys, the distribution of the vineyards among Rapel, Curicó, 253 and Maule was 35%, 20%, and 45%, respectively. 254 In terms of management strategies, 39% of the sample had pressurized systems to irrigate the vineyard 255 and 17% used machinery to perform the harvest. The tendone training system was a minority 256 compared to the vertical system (18% vs 82%, respectively), and only 11% of the growers produced 257 reserve quality grapes while the remaining 89% produced varietal quality. 258 As for climate-related conditions, the average potential evapotranspiration and precipitation of the 259 three warmest months in Chile, during the stage of veraison in grapes (period of accumulation of 260 sugars), were 456 mm and 23 mm, respectively. Concerning annual cumulative chilling hours, the 261 sample mean was 1,287 hours with a wide range (750 to 1,830 hours). 262 263 4.2. Contribution of production factors, vineyards’ attributes, management strategies and 264 climate-related conditions 265 As mentioned in Section 3, three sets of explanatory variables were progressively added to the basic 266 production function (Model A) to select the most appropriate specification to explain wine grape 267 growers’ TVP. Four specifications, one for each set of regressors, were estimated and compared 268 through maximum likelihood ratio tests. Table 2 reports the TVP model for the main variety of the 269 vineyard under the four alternative models. 270 271 Table 2. Cobb-Douglas estimates for total value product of Chilean wine grape growers under four 272 alternative models (N=336). 273 Model A: Model B: Model C: Model D: Production factors A + Vineyards’ attributes B + Management strategies C + Climatic conditions Variable Coeff. a Coeff.a Coeff.a Coeff.a Ln Land 0.603 *** 0.806 ** * 0.913 *** 0.917 *** Ln Fertilizers 0.033 0.018 0.018 0.020 WEP – Wine Economics and Policy Just Accepted Manuscript 11 Ln Fungicides 0.049 *** 0.028 ** 0.025 ** 0.022 ** Ln Agrochem 0.110 *** 0.066 ** 0.060 ** 0.054 ** Ln Labor 0.274 *** 0.156 ** * 0.056 0.050 Grape Color -0.381 ** * -0.384 *** -0.371 *** Vineyard age -0.163 ** * -0.112 *** -0.109 *** Rapel valley 0.262 ** * 0.246 *** 0.137 Maule valley -0.189 ** -0.168 ** -0.161 ** Irrig method 0.088 0.117 * Mech harvest -0.018 -0.019 Training system 0.492 *** 0.513 *** Grape Dest 0.227 ** 0.222 ** Ln Evapotransp 0.066 Ln Precipitation -0.275 ** Ln Chilling hours 0.123 Constant 1.394 *** 2.011 ** * 1.674 *** 1.246 Obs (N) 336 336 336 336 Adjusted R2 0.831 0.864 0.880 0.876 BIC 635.68 7 587.49 9 567.751 580.63 7 a Significance: ***=1%; **=5%; *=10%. First, model A – the basic production function including land, inputs, and labor – presents significant 274 parameters for all the covariates except for fertilizers. The base model was complemented with 275 covariates representing vineyards’ attributes (i.e., grape color, vine age, and wine valleys) resulting 276 in model B. To compare models A and B, a likelihood ratio test was performed to verify the 277 hypothesis that the former nested in the latter (i.e., additional covariates do not add to the explanation 278 of growers’ TVP). The test rejected the null hypothesis (p-value of 0.000 with 4 degrees of freedom), 279 giving support to the inclusion of vineyards’ attributes. Subsequently, we included the set of 280 management strategies (i.e., irrigation method, training system, mechanized harvest, and grape 281 destination) into model B to produce model C. The null hypothesis that model B is nested in model 282 WEP – Wine Economics and Policy Just Accepted Manuscript 12 C is rejected (p-value of 0.000 with 4 degrees of freedom), supporting the consideration of 283 management strategies in modelling growers’ TVP. Finally, climate-related variables (i.e., 284 evapotranspiration, precipitation, and chilling hours) were included in model C to produce model D. 285 The likelihood ratio test in this case did not favor model D (p-value of 0.207 with 3 degrees of 286 freedom), which explains that adding climate-related variables did not contribute to explaining 287 growers’ TVP. 288 In addition, we tested the inclusion of climate-related conditions in models A and B to corroborate 289 whether these variables have an effect in alternative models (results not shown but available upon 290 request). Only in model A was the inclusion of climate-related conditions supported by the likelihood 291 ratio test (p-value of 0.000 with 3 degrees of freedom), while in model B it was not (p-value of 0.704 292 with 3 degrees of freedom). Thus, the inclusion of climate-related variables into the TVP models was 293 not supported by statistical tests, except for the base model. Although somewhat unexpected, we 294 believe that there is a competing effect between climate-related conditions and the variables 295 controlling for vineyard location (i.e., the categorical variables for wine valleys). Indeed, analyses of 296 variance demonstrate statistically significant differences for the climate-related variables across 297 valleys (see Appendix 3). Each valley has distinct characteristics that are captured by the climate-298 related variables (for a further description of valley characteristics see Appendix 1). An additional 299 possible explanation for the non-significant effect of climate-related variables in model D is the date 300 of the primary and GIS data, which differed in one productive season. Specifically, the survey was 301 administered to grape growers in 2014-2015, and the environmental information from GIS referred 302 to 2015-2016. Although the timing of these two sources of information is not exact, due to GIS data 303 availability, climate-related variables in this study contribute to characterizing the microclimate of 304 the wine valleys included in the sample. 305 From the above, we can conclude that model C is preferred over the four confronted specifications, 306 being selected as the most appropriate to explain growers’ TVP. It should also be noted that goodness 307 of fit statistics reported at the bottom of Table 2 confirm that model C is the best alternative 308 (maximum Adjusted R-squared and lower Bayesian Information Criterion). Hence, model C is further 309 discussed in the following section. 310 311 4.3. Results and discussion of the Selected Model C 312 Table 2 shows that nine out of 13 covariates were significant (p<0.05) and explained 88% of the 313 variance of growers’ TVP. The estimated parameters must be interpreted as partial elasticities of 314 production (or percentage impact after exponentiating coefficients in the case of dummy covariates) 315 because of the logarithmic metric used in the model. The parameters of conventional inputs, here 316 WEP – Wine Economics and Policy Just Accepted Manuscript 13 referred to land, inputs, and labor, are all positive and less than one, and thus consistent with economic 317 theory [25]. The sum of these coefficients was 1.073, which was tested for constant return to scale. 318 The null hypothesis was rejected (p-value of 0.014 with 1 degree of freedom), hence we concluded 319 that the production function exhibits increasing returns-to-scale. This result is consistent with the 320 findings of Galindro et al. [18], who analyzed vineyard size in the Demarcated Douro Region of 321 Portugal, and with the findings of Sheng et al. [26] who found increasing returns to scale using a 322 sample of different agricultural establishments in Australia. 323 The parameter of the variable Land had a significant contribution in the explanation of growers’ TVP, 324 with an average elasticity of 0.91, meaning that a 10% increase in planted area translates into a 9.1% 325 higher TVP, when holding all other variables constant. Concerning other inputs, pesticides (i.e., 326 fungicides and other agrochemicals) were all significant, while fertilizers were not. These results may 327 be explained by the inherent characteristics of the crop (i.e., the Vitis genus), as wine grapes are highly 328 attractive to pests and diseases due to their elevated content of water and sugar, and vines have a 329 natural tendency to grow vigorously. Fertilization management, as in the case of irrigation, must be 330 carefully administered to the vineyard in order to have a correct balance between vegetative growth 331 and fruit production [27]. The latter seems to be supported by the data used in our study since 332 fertilizers, compared to pesticides, represent a smaller fraction in the total expenditure (sample 333 average sum of fungicides, insecticides, acaricides, herbicides, and fertilizers; see Table 1). The use 334 of fungicides increases the TVP with an average elasticity of 0.025 (i.e., a 10% increase in fungicide 335 expenditure translates into a 0.25% higher TVP). As for other agrochemicals – that includes 336 insecticides, acaricides, and herbicides – the growers’ TVP increases by 0.6% when the expenditure 337 in this item rises 10%. These results are expected since grapes are very sensitive to fungus, such as 338 powdery mildew, botrytis, and grapevine trunk diseases [28, 29, 30] and pests, such as Lobesia 339 botrana, Brevipalpus chilensis, Pseudococcidae spp. [31, 32, 33]. 340 Concerning labor expenditure, corresponding to the sum of expenses of performing the different 341 management activities evaluated in this study, the estimated parameter was not significant. This result 342 was unexpected since models A and B showed a significant contribution of labor expenditure in 343 explaining growers’ TVP. The only difference between these models and model C is that the latter 344 includes management strategy variables; therefore, it is likely that its inclusion has diluted the effect 345 of labor. Indeed, alternative training systems and grape destinations have implications in terms of the 346 use of labor (i.e., harvest, pruning/mooring, tipping of shoots, de-sprouting, canopy defoliation, 347 physical weed control, and other labors). For instance, the tendone training system imposes several 348 limitations for mechanizability [34], which translates into a greater dependence on manual labor. 349 Then, management strategies may act as confounding variables with labor expenditure. To illustrate 350 WEP – Wine Economics and Policy Just Accepted Manuscript 14 the differences in labor expenditure by training system and grape destination, Tables A.2 in Appendix 351 2 present a complete characterization of the vineyards, respectively. 352 As mentioned above, the training system and grape destination played a relevant role in our TVP 353 model, while pressurized systems and mechanized harvesting were not statistically significant. 354 According to our results, the training system is a determinant variable in the explanation of growers’ 355 TVP, increasing it by 63% when vineyards are trained as tendone compared to vertical training 356 systems (the marginal effect of binary variables correspond to their exponentiated parameter estimate 357 in model C). Grape destination was also significant in the model, showing that vineyards producing 358 reserve grapes (i.e., of superior quality) demonstrated a 25% increase in TVP compared to varietal 359 oriented vineyards. Appendix 2 show that tendone training systems exhibit considerably higher yields 360 and harvest expenditure and lower prevalence of mechanized harvesting and agrochemical 361 expenditure. The reserve quality grape destination, for its part, presents lower yields that are 362 compensated by higher prices to demonstrate a higher TVP (compared to varietal). As expected, it 363 also presents a higher aggregate labor expenditure (see item other labors). 364 As for vineyards’ attributes, all the variables included within this category were significant in 365 explaining growers’ TVP. It was found that vineyards growing red grape varieties generate 32% less 366 TVP than vineyards growing white grapes, holding all other variables constant. This is because white 367 grape varieties receive higher prices and present higher yields than red grape varieties in our sample: 368 the average price per kilo is USD$ 0.292 vs USD$ 0.246, respectively, and the average yield per 369 hectare is 16.7 tons and 14.5 tons, respectively. The age of the vineyard also plays a relevant role in 370 the model, indicating that TVP is reduced by 1.1% when the age is increased by 10%. In the empirical 371 literature there is mixed evidence on this topic, particularly on yield effects rather than on grape 372 quality effects. Some studies have found that vine age may reduce yields [35], while others have 373 found a positive [36] or no significant effect on yields [37]. 374 In terms of production valleys, using Curicó as a reference, wine grape growers from Rapel exhibit 375 28% higher TVP while those from Maule are 16% lower. That is to say, the growers’ TVP increases 376 as moving north in the study area. This result corresponds with average data displayed in Table A.3 377 (see Appendix 3), showing that growers from the northernmost valley (i.e., Rapel) present higher 378 average grape prices and yields. The same table shows that growers from Rapel face a lower incidence 379 of precipitation and higher evapotranspiration between December and February, which may affect 380 positively quality and yields, respectively. 381 382 4.4. Total value product functions derived from model C 383 WEP – Wine Economics and Policy Just Accepted Manuscript 15 Figure 2 displays several TVP functions for the production factors considered in this study (i.e., land, 384 fertilizers, fungicides, other agrochemicals, and labor) and the age of the vines. They represent the 385 relationship between each of these variables and vineyards’ outcomes, by showing the average 386 prediction of TVP in the sample (fitted value) at increasing values of the variable, holding all other 387 covariates in the model constant at observed values. In each TVP function, the pair of coordinates 388 that correspond with the median value of the variable (X-axis) and their expected TVP (Y-axis) is 389 presented. For example, in the case of land, the median value is 10 hectares, which is associated with 390 an expected TVP of US$ 29.854, holding all other covariates in the model constant at observed values 391 (see Figure 2.a). It can also be seen that there is a positive and almost linear (barely concave) response 392 of TVP as the quantity of hectares of vineyard increase. Notwithstanding, in the case of fertilizers, 393 fungicides, other agrochemicals, and labor, the concavity of the TVP function is very clear, which 394 indicates that the marginal effect of these variables is positive but decreasing. As for the age of vines, 395 the relationship is negative and convex, showing a decreasing marginal effect on TVP as the number 396 of years increase (see Figure 2.f). 397 a) b) c) d) WEP – Wine Economics and Policy Just Accepted Manuscript 16 e) f) Figure 2. Total value product functions from a sample of 336 Chilean wine grape growers for: a) 398 land, b) expenditure in fertilizers, c) expenditure in fungicides, d) expenditure in other agrochemicals, 399 e) expenditure in labor, and f) age of vines. In each graph there are plotted five data points that, from 400 left to right, correspond to the 10th, 25th, 50th, 75th and 90th percentiles. Therefore, coordinates (X, 401 Y) represent median values in X and the associated values in Y. 402 403 5. Conclusions and production implications 404 The economic analysis carried out in this study showed the impact of alternative management 405 strategies and cultural practices, controlling for vineyards’ structural variables and production 406 conditions, using a sample of 336 vineyards. Among significant variables, the results reveal that the 407 vineyard training system, grape color, grape destination, and vineyard age play an important role in 408 explaining growers’ total value product (TVP). In particular, a better economic performance is 409 expressed by vineyards using tendone training systems, growing white varieties, producing reserve 410 quality grapes, and having younger aged vines. These results have direct implications for both wine 411 grape growers and sectorial policy makers aiming to improve the competitiveness of viticultural 412 production by providing management strategies that result in better outcomes. In addition, we 413 improve on the existing literature as our results are based on a diverse, comprehensive, and relatively 414 large dataset, while previous studies tend to focus on specific or narrow factors of economic 415 performance (e.g., testing the effect of a particular management practice) and generally use purposive 416 samples that do not guarantee diversity or representativeness. In this regard, we disentangle the role 417 of a diversity of factors affecting viticultural production and estimate their impact on growers’ TVP, 418 which at the end is the ultimate goal of a vineyard. 419 We also included in the econometric model a set of climate-related variables from a GIS, which do 420 not appear to be significant in explaining growers’ TVP. This result was unexpected since agricultural 421 systems are naturally determined by climatic conditions, especially in recent years as they are 422 WEP – Wine Economics and Policy Just Accepted Manuscript 17 increasingly challenged by climate change. We believe that the joint inclusion of climate-related 423 variables in the econometric models with other crucial variables for wine grape growing (particularly, 424 the valley of production) competed in explaining the variance. In this regard, the study area of this 425 paper is centered in three important and traditional wine valleys of central Chile, the core of the 426 country’s vineyard production, which at some point capture climate-related conditions. The results 427 indicate that vineyards located in northern wine valleys – characterized by a lower on-season 428 precipitation, lower annual chilling hours, and higher evapotranspiration – demonstrate a higher 429 growers’ TVP. Another potential reason for the non-significant effect of climate-related variables, 430 apart from the competing effect by the variance with the valley of location in the statistical models, 431 is that vineyards are not as sensitive as other crops to the climate-related variables analyzed in this 432 paper. We suggest more research on this topic; deeper analyses are needed to explore this eventual 433 trait of vines as our data and analyses are limited in this regard. Future research might explore the 434 adaptive capacity of vines compared to other crops in light of the climate change phenomena affecting 435 our planet. 436 437 Despite the contributions of this paper, there were some inherent limitations that can be considered 438 by future investigations. First, in this study we use the main grape variety plot of the vineyard as the 439 unit of analysis, but it is likely that growers produce several grape varieties within a vineyard. Future 440 studies might consider this complexity when analysing economic performance by modelling 441 simultaneously the different outcomes of vineyards. Second, we believe that subsequent studies may 442 improve the findings presented here by including soil heterogeneity variables that may have an 443 important effect on vineyards’ economic performance. Although our model barely captured this effect 444 through the variable valley of location, we suggest the consideration of specific measures of the terroir 445 aiming to isolate this source of variability. Third, today’s digital technologies, such as GPS, PDA, 446 remote sensing or GIS, are becoming relevant in agricultural systems as they generate valuable 447 information to make better decisions and thus turn production processes more efficient. In our study, 448 we did not consider the adoption of these technologies as a management strategy that allows for 449 making precision agriculture at the sub-plot level. We acknowledge it as a shortcoming that could be 450 addressed in future research on this topic. 451 The main contribution of this paper is to advance in the understanding of economic performance 452 factors in wine grape growing, by simultaneously considering management strategies, production 453 conditions, and vineyards’ attributes. Capturing the effects of on-farm decisions made by the 454 vineyards, using a relatively large sample distributed in three different wine valleys, represents 455 valuable information to develop a strategy for the primary sector in Chile, which faces significant 456 WEP – Wine Economics and Policy Just Accepted Manuscript 18 competitiveness challenges compared to other agents of the marketing chain. Hence, our findings are 457 hopefully valid for other emergent countries in the global wine industry, and especially for those that 458 enjoy a Mediterranean climate. The practical implication of identifying what factors allow vineyards 459 to be more profitable serves to guide on-farm decisions of the private sector, both growers and 460 investors. Notwithstanding, the above is especially relevant for policy makers, to the extent that 461 improved economic performance at the vineyard level can have an aggregate impact on the 462 commercial success of the whole industry. 463 464 Acknowledgments 465 This study was financed by the National Fund for Scientific and Technological Development, 466 FONDECYT, project N° 1140615, from the National Commission for Scientific and Technological 467 Research, CONICYT, Chile. 468 469 Conflict of Interest 470 The authors declare no conflict of interest. 471 472 WEP – Wine Economics and Policy Just Accepted Manuscript 19 Appendices 473 474 Appendix 1 475 Valley Surveyed producers Characteristics Rapel 164 Composed by the sub-valleys Cachapoal and Colchagua, both are located in the O’Higgins region of Chile and are characterized by their sub-humid, Mediterranean temperate climate, ideal for the production of red varieties. The hours of light, high thermal oscillation, and the existence of various microclimates allow for growing different wine varieties. This region has a pronounced seasonality, where winter concentrates the most of annual rainfall. It has an average temperature of 22 °C and precipitation around 600 mm. The soils are alluvial in origin. These valleys are located north of the Curicó and Maule valleys. Curicó 91 Located in the Maule region of Chile, Curicó valley is considered the center of the Chilean wine growing because of its high concentration of vineyards. It has a temperate Mediterranean climate with a dry period five months a year, precipitation around 700 mm, and an average temperature of 20 °C. White varieties are best grown in the coolest areas of the valley. It has numerous water sources and the soil is alluvial and volcanic in origin. Maule 181 Located in the Maule region of Chile south of Curicó valley and considered the "Cradle of Chilean wine" because of its origin during the time of Spanish colonization. It has a temperate Mediterranean climate with rainy winters. The soils are acidic and clayed, which partially reduces productivity to benefit the quality of the grapes. It has many rivers that also exert influence on the quality of their wines. Total 436 476 477 478 479 480 481 WEP – Wine Economics and Policy Just Accepted Manuscript 20 Appendix 2 482 483 Table A.2. Vineyards’ characterization by training system and grape destination. 484 485 Training system Grape destination Vertical Tendone Varietal Reserve Variable N Mean N Mean N Mean N Mean Grape price (USD kg-1) 275 0.260 61 0.229 298 0.235 38 0.409 Yield (ton ha-1) 275 12.609 61 26.000 298 15.554 38 11.011 Planted area (ha) 275 17.297 61 14.249 298 16.644 38 17.527 Fertilizer expenditure (1,000 USD) 275 4.228 61 4.818 298 4.468 38 3.291 Fungicide expenditure (1,000 USD) 275 3.111 61 1.904 298 2.807 38 3.560 Expenditure in agrochemicals to control insects, spiders and weeds (1,000 USD) 275 6.453 61 3.883 298 5.674 38 8.435 Labor expenditure (1,000 USD) 275 15.680 61 20.116 298 16.226 38 18.521 Expenditure in pruning/mooring (1,000 USD) 270 4.616 61 7.181 295 5.174 36 4.392 WEP – Wine Economics and Policy Just Accepted Manuscript 21 Expenditure in harvesting (1,000 USD) 265 5.789 60 10.373 287 6.567 38 7.154 Expenditure in desprouting (1,000 USD) 232 1.722 47 1.355 247 1.645 32 1.777 Expenditure in thinning of shoots (1,000 USD) 217 0.895 26 0.489 214 0.858 29 0.808 Expenditure in physical weed control (1,000 USD) 200 0.985 52 0.953 229 0.971 23 1.048 Expenditure in other labors (1,000 USD) 167 4.436 27 1.508 167 3.665 27 6.276 Grape color (red=1; white=0) 275 0.829 61 0.803 298 0.829 38 0.789 Age of planting (years) 275 32.335 61 18.574 298 29.658 38 31.237 Irrigation method (pressurized= 1; gravity= 0) 275 0.378 61 0.459 298 0.396 38 0.368 Machinery use for harvest (yes= 1; no= 0) 275 0.200 61 0.033 298 0.178 38 0.105 Training system (tendone=1; vertical=0) 275 - 61 - 298 0.201 38 0.026 Grape destination (reserve=1; varietal=0) 275 0.135 61 0.016 298 - 38 - 486 487 488 WEP – Wine Economics and Policy Just Accepted Manuscript 22 Appendix 3 489 490 Table A.3. Mean comparison of grape price, yield and climate-related variables across valleys. 491 492 Variable Rapel Curicó Maule Grape Price (USD kg-1) 0.30 a 0.25 b 0.22 b Vineyard yield (ton ha-1) 17.42 a 15.22 a 12.63 b Precipitation (mm) 15.24 a 27.16 b 26.65 b Evapotranspiration (mm) 464.28 a 453.27 b 450.06 b Chilling hours (hours) 1009.13 a 1542.43 b 1395.87 c * Different letters within the same row means statistically significant differences (p< 0.05) 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 WEP – Wine Economics and Policy Just Accepted Manuscript 23 References 513 514 515 [1] Pappalardo, G., Scienza, A., Vindigni, G., D’Amico, M. (2013). Profitability of wine grape 516 growing in the EU member states. J. Wine Res. 24, 59–76. 517 https://doi.org/10.1080/09571264.2012.724392 518 [2] Newton, S.K., Gilinsky, A., Jordan, D. (2015). Differentiation strategies and winery financial 519 performance: An empirical investigation. Wine Econ. Policy 4, 88–97. 520 https://doi.org/10.1016/j.wep.2015.10.001 521 [3] Tudisca, S., Di Trapani, A.M., Sgroi, F., Testa, R. (2013). The cost advantage of sicilian wine 522 farms. Am. J. Appl. Sci. 10, 1529–1536. https://doi.org/10.3844/ajassp.2013.1529.1536 523 [4] Cusmano, L., Morrison, A., Rabellotti, R. (2010). Catching up trajectories in the wine sector: A 524 comparative study of Chile, Italy, and South Africa. World Dev. 38, 1588–1602. 525 https://doi.org/10.1016/j.worlddev.2010.05.002 526 [5] Sellers, R., Alampi-Sottini, V. (2016). The influence of size on winery performance: Evidence 527 from Italy. Wine Econ. Policy 5, 33–41. https://doi.org/10.1016/j.wep.2016.03.001 528 [6] Smith, K. (2007). Technological and economic dynamics of the world wine industry: An 529 introduction. Int. J. Technol. Glob. 3, 127–137. https://doi.org/10.1504/IJTG.2007.014329 530 [7] Anderson, K., Norman, D., Wittwer, G. (2001). Globalization and the World’s Wine Markets: 531 Overview. 532 [8] Castillo, J.S., Compés, R. (2013). The wine economy in Spain and the world. Cajamar Caja Rural. 533 [9] Fleming, E., Mounter, S., Grant, B., Griffith, G., Villano, R. (2014). The New World challenge: 534 Performance trends in wine production in major wine-exporting countries in the 2000s and their 535 implications for the Australian wine industry. Wine Economics and Policy, 3(2), 115-126. 536 [10] Santos, M., Rodríguez, X. A., Marta-Costa, A. (2021). Productive efficiency of wine grape 537 producers in the North of Portugal. Wine Economics and Policy, 10(2), 3-14. doi: 10.36253/wep-538 8977 539 [11] Moreira, V.H., Troncoso, J.L., Bravo-Ureta, B.E. (2011). Technical efficiency for a sample of 540 Chilean wine grape producers: A stochastic production frontier analysis. Cienc. e Investig. Agrar. 38, 541 321–329. https://doi.org/10.4067/S0718-16202011000300001 542 [12] Engler, A., Jara-Rojas, R., Bopp, C. (2016). Efficient use of water resources in vineyards: A 543 recursive joint estimation for the adoption of irrigation technology and scheduling. Water Resour. 544 Manag. 30. https://doi.org/10.1007/s11269-016-1493-5 545 WEP – Wine Economics and Policy Just Accepted Manuscript 24 [13] ODEPA (Office of Agricultural Studies and Policies). (2016). Foreign trade statistics: Advance 546 by product of exports/imports (In Spanish). Retrieved from: https://www.odepa.gob.cl/estadisticas-547 del-sector/comercio-exterior. Visited September 2021. 548 [14] FNE (National Economic Prosecutor's Office). (2017). “FNE rules out competitive abuses 549 against wine grape producers” (In Spanish). Retrieved from: https://www.fne.gob.cl/fne-descarta-550 abusos-competitivos-contra-productores-de-uva-vinifera-pero-detecta-otras-posibles-551 imperfecciones-en-este-mercado-y-las-informa-al-ministerio-de-agricultura-sag-y-odepa/. Visited 552 September 2021. 553 [15] Czyżewski, A., Smędzik-Ambroży, K. (2015). Specialization and diversification of agricultural 554 production in the light of sustainable development. J. Int. Stud. 8, 187–196. 555 https://doi.org/10.14254/2071-8330.2015/8- 556 [16] Keller, M. (2010). Managing grapevines to optimise fruit development in a challenging 557 environment: A climate change primer for viticulturists. Aust. J. Grape Wine Res. 16, 56–69. 558 https://doi.org/10.1111/j.1755-0238.2009.00077.x 559 [17] Manandhar, A., Zhu, H., Ozkan, E., Shah, A. (2020). Techno-economic impacts of using a laser-560 guided variable-rate spraying system to retrofit conventional constant-rate sprayers. Precis. Agric. 561 21, 1156–1171. https://doi.org/https://link.springer.com/article/10.1007/s11119-020-09712-8 562 [18] Galindro, A., Santos, M., Santos, C., Marta-Costa, A., Matias, J., Cerveira, A. (2018). Wine 563 productivity per farm size: A maximum entropy application. Wine Econ. Policy 7, 77–84. 564 https://doi.org/10.1016/j.wep.2018.03.001 565 [19] Urso, A., Timpanaro, G., Caracciolo, F., Cembalo, L. (2018). Efficiency analysis of Italian wine 566 producers. Wine Econ. Policy 7, 3–12. https://doi.org/10.1016/j.wep.2017.11.003 567 [20] Smart, R., Dick, J., Gravett, I, Fisher, B. (1990). Canopy management to improve grape yield 568 and wine quality - Principles and practices. S. Afr. J. Enol. Vitic. 11, No. 1. 569 [21] Keller, M. (2020). The Science of Grapevines. Third Edition, Academic Press. Elsevier 570 [22] Chilean Natural Resources Information Center (CIREN). Agroclimatic information (online). 571 Retrieved from: https://www.ciren.cl/productos/informacion-agroclimatica/. Visited by: September, 572 2020. 573 [23] Chinnici, G., Pecorino, B., Rizzo, M., Rapisarda, P. (2013). Evaluation of the performances of 574 wine producers in Sicily1. Calitatea 14, 108. 575 [24] StataCorp. (2017). Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC. 576 [25] Njuki, E., Bravo-Ureta, B.E. (2018). Irrigation water use and technical efficiencies: Accounting 577 for technological and environmental heterogeneity in U.S. agriculture using random parameters. 578 Water Resour. Econ. 24, 1–12. 579 WEP – Wine Economics and Policy Just Accepted Manuscript 25 [26] Sheng, Y., Zhao, S., Nossal, K., Zhang, D. (2015). Productivity and farm size in Australian 580 agriculture: reinvestigating the returns to scale. Australian Journal of Agricultural and Resource 581 Economics, 59(1), 16-38. 582 [27] Maughan, T., Pace, M., Black, B. (2017). Grape vine management. 583 [28] Valdés‐Gómez, H., Araya‐Alman, M., Pañitrur‐De la Fuente, C., Verdugo‐Vásquez, N., Lolas, 584 M., Acevedo‐Opazo, C., Gary, C., Calonnec, A. (2017). Evaluation of a decision support strategy for 585 the control of powdery mildew (Erysiphe necator [Schw.] Burr.), in grapevine in the central region 586 of Chile. Pest management science, 73(9), 1813-1821. 587 [29] Pañitrur-De La Fuente, C., Valdes-Gómez, H., Roudet, J., Acevedo Opazo, C., Verdugo-588 Vasquez, N., Araya-Alman, M., Lolas, M., Moreno, Y., Fermaud, M. (2018). Classification of 589 winegrape cultivars in Chile and France according to their susceptibility to Botrytis cinerea related 590 to fruit maturity. Australian Journal of Grape and Wine Research 24(2):145-157. 591 [30] Silva-Valderrama I., Toapanta D., Miccono M., Lolas M., Díaz G., Cantu, D., Castro A. (2021). 592 Biocontrol potential of grapevine endophytic and rhizospheric fungi against trunk pathogens. Front. 593 Microbiol. 11:614620. doi: 10.3389/fmicb.2020.614620 594 [31] Altimira, F., Vitta, N., Godoy, P., Tapia, E. (2020). Microbial pesticides for the integrated 595 management of Lobesia botrana in grapevines. Newsletter INIA Nº 419. Agricultural Research 596 Institute (INIA). Regional Research Center La Platina, La Pintana, Chile. 96 p. 597 [32] Curkovic, T., Durán, M., Ferrera, C. (2013). Control of Brevipalpus chilensis Baker (Acari: 598 Tenuipalpidae) with agricultural detergents under laboratory and field conditions. Chilean Journal of 599 Agricultural & Animal Sciences, ex Agro-Ciencia. 29 (1), 73-82. 600 [33] Sazo, L., Araya, J., de la Cerda, J. (2008). Effect of a siliconate coadjuvant and insecticides in 601 the control of mealybug of grapevines, Pseudococcus viburni (Hemiptera: Pseudococcidae). 602 International Journal of Agriculture and Natural Resources, 35(2), 215-222. 603 https://dx.doi.org/10.4067/S0718-16202008000200012 604 [34] Cogato, A., Pezzuolo, A., Sørensen, C., De Bei, R., Sozzi, M., Marinello, F. (2020). A GIS-605 based multicriteria index to evaluate the mechanisability potential of Italian vineyard area. Land, 606 9(11), 469. 607 [35] White, R. (2003). Soils for fine wines. Oxford University Press. 608 [36] Grigg, D., Methven, D., De Bei, R., Rodríguez, C., Dry, P., Collins, C. (2018). Effect of vine 609 age on vine performance of Shiraz in the Barossa Valley, Australia. Australian journal of grape and 610 wine research, 24(1), 75-87. 611 [37] Considine, J. (2004). Grapevine productivity and yield components: A case study using field 612 vines of Zante currant. Australian journal of grape and wine research, 10(2), 108-115. 613