Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 APPLICATION OF THE STANDARDIZED VEGETATION INDEX (SVI) AND GOOGLE EARTH ENGINE (GEE) FOR DROUGHT MANAGEMENT IN PERU † [APLICACIÓN DEL ÍNDICE DE VEGETACIÓN ESTANDARIZADO (SVI) Y GOOGLE EARTH ENGINE (GEE) PARA LA GESTIÓN DE SEQUÍAS EN PERÚ] Jaris E. Veneros*1,2 and Ligia García2 1Department of Ecology, Montana State University, 1156-1174 S 11th Ave, Bozeman, Montana 59715, USA. 2Instituto de Investigación para el Desarrollo Sustentable de Ceja de Selva (INDES-CES.).Universidad Nacional Toribio Rodríguez de Mendoza de Amazonas. 342 Higos Urco, Chachapoyas 01001, Perú. Email: jaris.veneros@untrm.edu.pe *Corresponding author SUMMARY Background. The SVI (Standardized Vegetation Index) provides a relative comparison of the condition of the vegetation in different classifications for monitoring droughts. Objective. In this research, the SVI was used through the Google Earth Engine (GEE) at the national level and in three study points for a coastal, Amazonian, and Andean region for October 31, 2020, and two decades. Methodology. For the construction of the SVI, the data from the Moderate Resolution Imaging Spectroradiometer (MODIS) Version 6 were used; of the Terra sensor (MOD13Q1) with a temporal resolution of 16 days, a spatial resolution of 250 meters, and as a level 3 product. Results. The SVI was represented in five classifications: with green color ≥ 0 (No Drought), yellow color -0.10 to -0.94 (Slight drought), light orange color -0.95 to -1.44 (Moderate drought), dark orange color -1.45 to -1.94 (Severe drought), and red color ≤ -1.95 (Extreme drought). Implications. The change in historical SVI values was evidenced due to causes such as El Niño costero (coastal) and deforestation of Tropical Forests; for the Sechura Desert in Piura and La Pampa in Madre de Dios, respectively. Subsequently, in the Andes of Peru, in Ollachea, Puno, it was determined that the SVI value, more extreme negative, represented an extreme drought never registered for this area. Conclusion. The SVI and GEE provided tools for drought management with high spatial and temporal resolution. Keywords: SVI; drought; vegetation condition; Google Earth Engine. RESUMEN Antecedentes. El SVI (Índice de Vegetación Estandarizado) proporciona una comparación relativa del estado de la vegetación en diferentes clasificaciones para el seguimiento de las sequías. Objetivo. En esta investigación se utilizó el SVI a través del Google Earth Engine (GEE) a nivel nacional y en tres puntos de estudio para una región costera, amazónica y andina para el 31 de octubre de 2020 y dos décadas. Metodología. Para la construcción del SVI se utilizaron los datos del espectrorradiómetro de imágenes de resolución moderada (MODIS) versión 6; del sensor Terra (MOD13Q1) con una resolución temporal de 16 días, una resolución espacial de 250 metros y como producto de nivel 3. Resultados. El IVS se representó en cinco clasificaciones: con color verde ≥ 0 (Sin sequía), color amarillo -0,10 a -0,94 (Sequía leve), color naranja claro -0,95 a -1,44 (Sequía moderada), color naranja oscuro -1,45 a -1,94 (Sequía severa), y color rojo ≤ -1,95 (Sequía extrema). Implicaciones. Se evidenció el cambio en los valores históricos del SVI debido a causas como El Niño costero y la deforestación de los Bosques Tropicales; para el Desierto de Sechura en Piura y La Pampa en Madre de Dios, respectivamente. Posteriormente, en los Andes del Perú, en Ollachea, Puno, se determinó que el valor del SVI, más extremo negativo, representó una sequía extrema nunca registrada para esta zona. Conclusión. El SVI y GEE proporcionaron herramientas para la gestión de la sequía con alta resolución espacial y temporal. Palabras clave: SVI; sequía; estado de la vegetación; Google Earth Engine. † Submitted May 19, 2021 – Accepted October 25, 2021. This work is licensed under a CC-BY 4.0 International License. ISSN: 1870-0462. Note: The current version updated the NDVI formula (page 3) on 9/11/2022. 1 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 INTRODUCTION Otherwise, understanding the relationship between vegetation vigor and moisture NASA in communication on October 26, 2020, availability is complex and has not been said that there is a severe drought in South adequately studied with satellite sensor data (Ji America, the first signs of the magnitude of the and Peters, 2003). Drought is a prolonged drought appeared in satellite gravimetry absence of rainfall, which generates water observations of southeastern Brazil since mid- scarcity over a sufficiently long period so that the 2018 and had spread to parts of Paraguay, lack of precipitation generates a serious Bolivia, and northern Argentina by 2020 (NASA, hydrological imbalance. Three types of droughts 2020), considered the second most intense have been determined (Zargar et al., 2011): a. drought in South America since 2002; this alert Agricultural drought is the deficit of humidity in is based on the extent, duration and volume of the upper meter of the soil, that is, in the root water lost during the drought as measured by the zone, which affects the crops (Skakun et al., GRACE and GRACE-FO satellites (Rodell, 2016). b. Meteorological drought, which is due 2020). A second announcement alerted the to prolonged precipitation deficit (Spinoni et al., scientific community on November 25, 2020, 2019), and c. Hydrological drought, which is when the Peruvian National Meteorological and related to below-normal groundwater, lake, and Hydrological Service (SENAMHI) reported that streamflow levels (Shamshirband et al., 2020). the deficiency of rainfall at the national level will Drought studies should consider the duration, remain of the same magnitude until the following magnitude, intensity, severity, geographic extent, weeks, especially in the central and southern and frequency of droughts (Zargar et al., 2011): highlands of this country, due to the low a. Duration, depending on the region, the humidity in the area (SENAMHI, 2020a). duration of drought can vary from one week to a Similarly, rainfall deficiencies at the national few years. Due to the dynamic nature of drought, level are associated with the entry of dry air from a region may experience episodes of drought and the Pacific, with greater incidence in the southern rainfall simultaneously when considering various highlands of Peru (SENAMHI, 2020b; time scales (Kibret et al., 2020). b. Magnitude, SENAMHI, 2020c), this dry air brought the the cumulative water deficit; e.g., precipitation, withdrawal of humidity from the Andes to the soil moisture, or runoff below a certain threshold east, favoring clear skies and intense radiation during a period of drought (Zargar et al., 2011). during the day and lower temperatures during the c. Intensity, the relationship between the night in the central and southern highlands magnitude of the drought and its duration (Kibret (Fernández, 2020). The humidity forecast for the et al., 2020). d. Severity, two uses are provided southern region for November 27 and 28 in 2020 for drought severity; the magnitude of the favored the occurrence of localized rainfall in the precipitation deficit, i.e., the magnitude and eastern region and less intensity in the western degree of impacts resulting from the deficit region, with the possibility of lesser and isolated (Xiangtao et al., 2020). e. Geographic extent, the rains until December of this year (SENAMHI, area coverage of the drought that is variable 2020a). Rainfall deficiencies persisted, during the event (Ghazaryan et al., 2020); this especially in the western regions of the Southern area can cover several kilometers (Skakun et al., Andes. The agricultural sector is affected by 2016) and f. Frequency or return period, defined increased water stress in Andean dryland crops, as the average time between drought events as well as by the delay in planting for November having a severity equal to or greater than a in the central and southern highlands of Peru threshold (Zargar et al., 2011). (Fernández, 2020). The impacts caused by SARS-CoV-2 on the agricultural sector, Drought monitoring relies on data from affecting the agri-food chain, added to this geostationary and polar-orbiting satellites, as problem (García et al., 2020). well as information in situ (Chuvieco, 2008). Besides, satellite information can be The Presidency of the Council of Ministers of complemented with studies related to Peru on December 1, 2020, issued Supreme groundwater levels to replenish lakes and Decree No. 149-2020-PCM declaring the State of reservoirs (Kibret et al., 2020). For example, Emergency for 60 calendar days in 38 provinces more than 74 drought indices are known (Zargar and 181 districts of the regions of Tumbes, Piura, et al., 2011), most of them derived from SPI Lambayeque, La Libertad, Cajamarca, and (Standardized Precipitation Indices) and NDVI Ancash, due to imminent danger of water deficit (Normalized Difference Vegetation Index) in the northern part of Peru (SENAMHI, 2020a), (Wainwright et al., 2020). There are also index this meant the transfer of functions to the depending on the purpose of the study; for regional governments in the implementation of example, detection and monitoring of droughts in the actions after technical studies, without real-time (Ghazaryan et al., 2020), declaration of specifying budget amounts (PCM, 2020). the beginning or the end of a drought period 2 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 (Tarnavsky and Bonifacio, 2020), a study of Lavado, 2018), in this sense., the launch of GEE, drought levels and drought response measures a cloud computing platform for storing and (Zargar et al., 2011), analysis of quantitative processing geographic datasets from local to impacts of droughts on variables at geographic planetary geographic scale (Gorelick et al., 2017; and temporal scales (Tarnavsky and Bonifacio, Kibret et al., 2020), helps in long-term drought 2020), finally the declaration of drought monitoring because of the tools it provides conditions among researchers, technicians, (Wang et al., 2012), and for being an important organizations and the general public (Adedeji et component in early warning systems. Therefore, al., 2020; Tsakiris and Vangelis, 2004; Zargar et the objective of this research was to analyze al., 2011). These indices currently assist in a droughts using the Standardized Vegetation variety of operations, such as drought warning, Index (SVI) and the GEE for two decades (2000 monitoring, and contingency planning. - 2020) in Peru and to expose their potential for planning and response to drought impacts at local The reports of Peru's National Center for the and national scales with a time scale of 16 days. Estimation, Prevention and Reduction of Disaster Risk (CENEPRED) based on scientific METHODOLOGY information from the National Hydrology and Meteorology Service (SENAMHI) and the Standardized Vegetation Index (SVI) Multisectoral Committee in charge of the National Study of El Niño costero (ENFEN) The SVI is used for drought monitoring and early regarding droughts or hostile events, are prepared warning of droughts. This index describes the on a regional, provincial or basin scale; however, probability of variation of the normal NDVI over this information is provided long after the event several years of data, in a weekly time interval and therefore hinders the response of decision (Peters et al., 2002). SVI is a z-score deviation makers in the process of estimating, preventing from the mean in units of standard deviation, and reducing disaster risk (CENEPRED, 2020). calculated from the NDVI (Normalized Difference Vegetation Index) and EVI Moderate Resolution Imaging Spectroradiometer (Enhanced Vegetation Index) values for each (MODIS) Version 6; Terra sensor (MOD13Q1) pixel location of a composite period for each year data are generated every 16 days at a spatial during a given reference period. The SVI formula resolution of 250 meters as a level 3 product is shown below (UN-SPIDER, 2020): (Ezzine et al., 2014; Didan, 2020), and are available free of charge at Google Earth Engine (GEE) (Da Silva et al., 2020). The SVI (Standardized Vegetation Index) is based on the calculation of Z-scores, a deviation from the mean NDVI in standardized deviation units at the Where: Zijk is the z-value for pixel i during week level of each pixel over a time series (Peters et j for year k, VIij is the weekly VI (Vegetation al., 2002). The SVI allows the time series to be Index) value for pixel i during week j for year k, extended with data from the National Oceanic so both NDVI and EVI (Son et al., 2014) can be and Atmospheric Administration Advanced Very used as VI, µij is the mean for pixel i during week High-Resolution Radiometer (NOAA-AVHRR) j for n years, and σij is the standardized deviation as long as a full inter-calibration between the two of pixel i during week j for n years. This formula sensor systems is provided (Swain et al., 2011). was established to obtain data every week; The SVI provides a relative comparison of however, due to the temporal resolution of the vegetation conditions, whereas the assessed satellite used in this study, measurements are deviation from the mean vegetation condition being taken every 16 days. cannot be translated into an absolute deviation of plant height, for example. Nor can the SVI be For the NDVI and EVI calculations (Son et al., interpreted as an absolute quantification of 2014): agricultural damage (Ezzine et al., 2014; Ji and Peters, 2003; Swain et al., 2011). NDVI = (Pnir - Pred)/(Pnir + Pred) Understanding the temporal and spatial behavior EVI = (2.5*Pnir - Pred)/(Pnir + 6*Pred - 7.5*Pblue + of precipitation is of high interest, especially in 1’) climate risk studies, where the availability of high resolution and good quality information is Where: Pred: (620–670 nm), Pnir (841–876 nm), essential (Carbajal et al., 2010). However, and Pblue (459–479 nm) are MODIS bands 1, 2, conventional rain gauge measurements are and 3 relatively scarce and poorly distributed, especially in developing countries (Asurza and 3 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 The SVI calculation is based on the EVI, which in cooperation with UN-SPIDER within the in turn is obtained from the corrected NDVI, has SEWS-D project (UN-SPIDER, 2020); MODIS improved sensitivity in dense vegetation images were used: ee.ImageCollection conditions, and is less affected by the influence ("MODIS/006/MOD13Q1") from GEE (Didan et of aerosols (Wang et al., 2012). The random al., 2015), from October 31, 2020. variable of a standardized normal distribution corresponds to a Z-score. Therefore; each random variable X can be transformed into a Z- Table 1. Drought classes according to the score by the following equation (UN-SPIDER, value of the Standardized Vegetation Index 2020): (SVI) for Peru Class Value Color Z = (X- μ)/σ Extreme Drought ≤ -1.95 Severe Drought -1.45 a -1.94 Where X is a normal random variable, μ is the Moderate Drought -0.95 a -1.44 mean and σ is the standard deviation. Therefore, Mild Drought -0.10 a -0.94 a Z-score equal to 0 represents an element equal No Drought ≥ 0 to the mean, a Z-score less than 0 represents an element less than the mean, and a Z-score greater than 0 represents an element greater than the Historical SVI mean. The Z-score indicates how many standard deviations an item is from the mean, so the The historical values and trends of SVI recorded standard deviation, in general, indicates how in two decades (2000-2020), were obtained with dispersed the data set is (Peters et al., 2002). A a spatial resolution of 250 m and a temporal low standard deviation implies that the data are resolution of 16 days, applying as examples of tightly clustered around the mean, while a high cases of one pixel per Region: Coastal (Sechura standard deviation implies that the data are Desert, Piura), Andean (Ollachea, Puno) and spread over a wider range of values. If the Amazonian (La Pampa, Madre de Dios) zones. number of elements in the data set is large, approximately 68% of the data are within 1 Terra-Modis vegetation indices standard deviation from the mean, 95% within 2 standard deviations, and 99.7% within 3 standard The MOD13Q1 V6 product available from GEE deviations from the mean, when it is a normal provides a Vegetation Index (VI) value per pixel distribution. The SVI was represented in five of 250 m, where there are two vegetation layers: categories (Table 1) (UN-SPIDER, 2020): SVI the first is the NDVI and the second vegetation with green color ≥ 0 (No Drought), yellow color layer is the EVI which minimizes canopy -0.10 to -0.94 (Mild Drought), light orange color background variations and maintains sensitivity -0.95 to -1.44 (Moderate Drought), dark orange in dense vegetation conditions (Didan et al., color -1.45 to -1.94 (Severe Drought) and red 2015). EVI also uses the blue band to remove color ≤ -1.95 (Extreme Drought). The SVI residual air pollution caused by smoke and thin dynamics can be influenced by rainfall, stress, sub-pixel clouds (Ronchetti et al., 2020). phenology, flooding, pests and diseases, nutrient MODIS products such as NDVI and EVI are deficiencies, forest fires, grazing and human calculated from atmospherically corrected activities (Ji Peters, 2003). However, the factors bidirectional surface reflectance that has been mentioned above must be associated with the masked for water, clouds, heavy aerosols, and types of droughts as well as the duration, cloud shadows (UN-SPIDER, 2020). magnitude, intensity, severity, geographic extent, and frequency to understand the droughts (Zargar Study area et al., 2011). Finally, we used the coefficient of variation to determine the statistical measure of The Republic of Peru is a country located in the dispersion of the data points around the mean South America and has a total area of 1 285 215 of the calculated SVI and box plots to show the km². t borders the Pacific Ocean to the west, dispersion of the SVI data in the study areas. Ecuador and Colombia to the north, Brazil to the These box plots are a standardized method of east, Bolivia to the southeast, and Chile to the graphically representing a series of SVI south. For this research, we used the shapefiles of numerical data through their quartiles. Peru at the Adm0 and Adm1 levels of the GADM Portal version 2.8 (GADM, 2018). The Andes SVI Nacional Mountain Range determines different geomorphologic units of a continental and The Standardized Vegetation Index (SVI) was marine environment for Peru (González- used for drought analysis at the national level Moradas Viveen, 2020). From west to east, in the using the method developed by UFSM in Brazil continental area, the units correspond to 1. 4 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Coastal Cordillera; 2. Pre-Andean Plain; 3. introduced by the Andes Mountains, and the flow Western Cordillera; 4. Inter-Andean along the country's coasts (INGEMMET, 1995). Depressions; 5. Eastern Cordillera; 6. Titicaca Basin; 7. Sub-Andean Region; and 8. Amazonian Sechura Desert, Piura - Coast Region Plain and in the marine field, the units include 1. Continental Shelf; 2. Continental Slope; 3. A random point of geographic coordinates was Marine Trench; 4. Nazca Ridge; and 5. South selected -5.315960, -81.034303 using the ArcGis Pacific Abyssal Sea Floor (INGEMMET, 1995). ver. 10.7 tool called Arc Toolbox: Data Management Tools > Feature Class > Create Concerning the climate, Peru has eight natural Random Points (Figures 1A, 2A y 3A). It is regions (Pulgar, 2014): Chala or coast, yunga, located in the Sechura Desert, in the Piura quechua, suni, puna, janca or mountain range, Region, with an average altitude of 11 masl, an high jungle, and low jungle. Therefore, Peru has average annual maximum temperature of 30 °C, a diversity of climates and microclimates from an average annual minimum temperature of 23 the arid and warm coastal, through the inter- °C and an average annual rainfall of 16 mm Andean valleys of temperate, frigid, and polar (SENAMHI, 2020b). This point presents an type to the warm and rainy type of the jungle ecosystem of the coastal desert type, i.e., arid to (SENAMHI, 2020b). Three factors determine hyper-arid climate with areas mostly devoid of Peru's climate: the country's location in the vegetation consisting of sandy soils or rocky intertropical zone, the altitudinal changes outcrops in flat, undulating, and dissected areas subject to wind erosion (MINAM, 2018). Figure 1. Standardized Vegetation Index (SVI) in Peru. 5 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 La Pampa, Madre de Dios - Jungle Region RESULTS AND DISCUSSIONS In the Madre de Dios Region, a random point was Regional Standardized Vegetation Index established with the following geographic (SVI) coordinates -12.853405, -69.972633 (Figures 1B, 2B, and 3B), this point belongs to the place It was determined that, as of October 31, 2020, called La Pampa. It has an average altitude of 203 SVI values in all regions of Peru presented masl, an average maximum temperature of 31 °C, ranges from extreme drought values (≤ -1.95; red an average minimum temperature of 20 °C and color) to non-drought values (≥ 0; green color) shows average annual precipitation of 2221 mm (Table 1, Figure 1). (Figures 1B, 2B and 3B) (SENAMHI, 2020b). This point is within an area dedicated to illegal SVI dynamics can be influenced by precipitation, mining and belongs to a buffer zone of the stress, phenology, flooding, pests and diseases, Tambopata National Reserve. The ecosystem is nutrient deficiency, forest fires, grazing, and an alluvial landscape in the Amazonian plain, human activities (Ji Peters, 2003). The which is periodically flooded by normal floods of coefficients of variation of the SVI range from 5 to 8 meters in height. The forest with sparse or 124.08% in Arequipa to 1691.75% in Puno open undergrowth can have 3 or 4 strata with a (Table 2, Figure 1); a situation that merits that the canopy or dome of trees that reach 20 to 25 analysis of the drought indexes should be meters high and emergent individuals up to 30 performed independently for each region, as well meters high (MINAM, 2018). as the meteorological conditions, especially if the seasons of the year are marked, such as autumn Ollachea, Puno - Andes Region and winter (Table 2, Figure 1) (Ezzine et al., 2014), because the most common statistical In the Puno Region, a random point was selected methods applied to NDVI and precipitation time with the geographic coordinates -13.792852, - series, such as simple linear correlation or 70.453065 (Figures 1C, 2C, and 3C), which regression analyses, produce inaccurate results if belongs to the Ollachea District. It has an average seasonality is not taken into account (Ji Peters, altitude of 2659 masl, an average maximum 2003). temperature of 22 °C, an average minimum temperature of 7 °C, and average annual The values for extreme droughts (minimum SVI, precipitation of 1360 mm (Figures 1C, 2C, and ≤ -1.95) exceed up to 2.16 times the classification 3C) (SENAMHI, 2020b). This point presents an limit value (UN-SPIDER, 2020), as in the case of ecosystem of the Pajonal type of humid puna the Ayacucho Region (-4.23), or 1.57 times as in (MINAM, 2018). the case of the Tumbes Region (-3.08) (Table 2; Figure 1); this showed that these places deserve special attention due to their low water availability. Recalling that, in addition to this Figure 2. Historical values for the Standardized Vegetation Index (SVI) for three pixels every 16 days from February 18, 2000 to October 31, 2020. Where: A. Sechura Desert, Piura B. La Pampa, Madre de Dios, and C. Ollachea, Puno. 6 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 A B C Figure 3. Trends of historical values for the Standardized Vegetation Index (SVI) from February 18, 2000 to October 31, 2020. Where: A. Sechura Desert, Piura B. La Pampa, Madre de Dios, and C. Ollachea, Puno. problem, there are the impacts caused by SARS- In Peru, SENAMHI uses interpolated CoV-2 on the Peruvian agricultural sector, in climatological and hydrological data with a terms of quantity and the need to distribute food resolution of 5 km to study rainfall and droughts for food security (García et al., 2020). (Risco Sence et al., 2017), but drought analysis based on Terra Modis data identifies patterns of drought-prone areas (Uttaruk Laosuwan, 2019) 7 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 at a resolution of 250 m (Didan, 2020), which management of deforestation monitoring. provides more timely data for technical studies Therefore; negative values in the SVI may be due on drought management, as requested by the to deforestation and the presence of bare soil. central government in its decree on drought management N° 149-2020-PCM (PCM, 2020). Ollachea, Puno Region Adjusting the SVI data obtained from satellites with SENAMHI observations may be of At the point under study, with geographical potential use for drought studies, depending on coordinates, -13.792852, -70.453065 (Figure the availability of data from the national network 1C), Ollachea district, Puno Region, presented a of meteorological stations in Peru. minimum SVI for the two decades of -3.97 (extreme drought) (Table 1) with a coefficient of Sechura Desert, Piura Region variation 1691.75 % (Table 2). The SVI presents extreme negative outliers for 2020 (Figures 2C The study point with geographical coordinates - and 3C); the most extreme value -3.97; was 5.315960, -81.034303 (Figure 1A), in the recorded for August 12, 2020, followed by other Sechura Desert, Piura Region, presented a extreme values observed in August and minimum SVI of -3.34 corresponding to extreme September, -3.10 and -3.66 respectively, all these drought (Table 1); with a coefficient of variation values correspond to Extreme Drought (Table 1). of 360.67 % in the two decades (Table 2). This According to a SENAMHI report, the Puno point is in the Sechura Desert, a pixel with mostly Region had a 95% rainfall deficit for November, negative values due to the natural conditions of and since October 24, 2020, rainfall was its ecosystem. Figures 2A and 3A, show that in insignificant in most of the stations of the Puno April 2017 there was an increase of SVI up to 3.5 highlands (SENAMHI, 2020d). (No drought), the atypical value within the two decades and represented by the appearance of Drought management should be done at a high vegetation conditions according to the SVI spatial scale and high geographic extent, interpretation. This is consistent with the rains identifying water sources, water balances and that occurred on those dates when it was declared examining downstream river flows (Gutiérrez et an area impacted by El Niño costero (coastal) al., 2005; Phillips et al., 2009), droughts are more (Salazar et al., 2019). These precipitations recurrent due to the lack of focus on the generated the greening of some parts of the protection of water sources (Garcia and Otto, Sechura Desert, according to NASA Worldview 2015). Countries such as Brazil used the ESG to satellite views. Finally, the most intense rains in support restoration policy measures within Piura occurred on March 26, with a duration of environmental sciences, territorial planning, and 15 hours of rain, so that on March 27, the Piura subsidy for the enforcement of environmental River overflowed its banks with a flow of 3,016 law (C. Silva Junior et al., 2020). In this sense, it m3/s, flooding Piura, Castilla, and agricultural has been identified that the origins of the areas (Salazar et al., 2019; Villa Aurelien, 2020). Amazon River area in the Peruvian Andes (Anderson et al., 2018), and drought La Pampa, Madre de Dios Region management in emergencies should be declared at the source, not only downstream, as has Droughts should be evaluated not only by the happened with rivers in the Pacific, Atlantic or presence or absence of rainfall but also by Titicaca basins (Lavado-Casimiro and Espinoza, surrounding factors that allow identifying, for 2014). example, regeneration patterns after fires (Chuvieco et al., 2020), adding studies of The actions taken by the Peruvian Government in geographic variables such as topography, soil the face of drought emergencies are oriented characteristics, or climate or relevance of burn solely to the purchase of products, for example, severity. For example, in La Pampa, Madre de water supplies for livestock, foliar fertilizers, Dios Region, for the study point with geographic supplementary feed, among others in the coordinates -12.853405, -69.972633 (Figure 1B), Cajamarca Region (COEN, 2020), historical values were identified for the two comprehensive drought mitigation strategies are decades of SVI with extreme droughts urgently needed. (maximum SVI: -3.62) (Figures 2B and 3B), corroborating the loss of tropical forests to bare Finally, the SVI was obtained from a deviation of soils, where illegal mining activities have been the z-score (pixel) from the mean in standard deforesting since before 1999 (Figures 2B and deviation units, calculated from NDVI and EVI 3B) (Asner Tupayachi, 2017), The negative values (Peters et al., 2002) in the Google Earth values of this SVI corroborate the lack of Engine (GEE), at the national level and historical vegetation conditions for this zone, so it is trends in Peru for two decades. The SVI suggested that this index complements the presented advantages concerning meteorological 8 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Table 2. Values for the Standardized Vegetation Index (SVI) for October 31, 2020 in the Regions of Peru Region N SVIminimum SVImaximum SVIaverage SDSVI CVSVI % Amazonas 78610 -3.62 3.50 -0.39 1.10 281.49 Ancash 107349 -3.81 4.28 -0.07 0.83 1139.23 Apurímac 64760 -3.62 3.82 -0.18 0.88 489.96 Arequipa 198522 -4.07 4.37 0.94 1.16 124.08 Ayacucho 116550 -4.23 4.36 0.10 1.12 1107.22 Cajamarca 100991 -3.61 3.14 -0.44 0.90 205.35 Cusco 195687 -3.92 3.94 0.24 0.82 345.78 Huancavelica 60244 -3.34 3.77 -0.21 0.85 402.04 Huánuco 94126 -3.18 3.70 -0.20 0.89 440.11 Ica 55308 -4.18 4.47 0.82 1.11 135.18 Junín 124599 -3.23 3.33 0.07 0.95 1291.16 La Libertad 85503 -3.47 4.45 -0.08 0.92 1104.44 Lambayeque 46092 -3.43 4.36 -0.09 0.85 969.85 Lima y Callao 102750 -3.78 4.32 0.14 0.93 655.13 Loreto 955497 -3.62 3.69 0.14 0.94 661.04 Madre de Dios 170387 -3.37 2.85 -0.12 0.98 802.63 Moquegua 41178 -4.11 4.31 1.19 1.16 96.82 Pasco 70765 -3.22 3.18 0.19 0.93 488.83 Piura 119186 -3.34 4.12 -0.22 0.78 360.67 Puno 184763 -3.97 4.04 0.05 0.86 1691.75 San Martín 104387 -3.71 3.43 -0.41 0.92 226.61 Tacna 46115 -4.01 4.34 1.09 1.09 99.39 Tumbes 15791 -3.08 2.69 -0.69 0.73 105.82 Ucayali 350644 -3.74 3.44 0.35 0.75 215.42 N=Number of pixels per Region; SVIminimum=Lowest SVI value by region; SVImaximum= Highest SVI value by region; SVIaverage=Average SVI value by region; SDSVI= Standard deviation for the SVI by region and CVSVI %= Coefficient of variation of SVI by region. drought indices (Ji and Peters, 2003). It was historical series of satellite images from 2000 to possible to work with a high spatial resolution at date, at the local and national level with a spatial the national level (250 m) and to cover larger scale of 250 m and a temporal scale of 16 days areas (Gorelick et al., 2017). A temporal for drought studies for Peru. resolution of 16 days allows better management to face droughts in Peru; achieving values for the Acknowledgments severity of vegetation stress resulting from a Thanks to the Instituto de Investigación para el water deficit (Wang et al., 2012) or other Desarrollo Sustentable de Ceja de Selva phenomena that cause the loss of water content (INDES-CES) of the Universidad Nacional in plants. Toribio Rodríguez de Mendoza of Amazonas (Peru) and Montana State University (USA) for CONCLUSIONS the academic support. The Standardized Vegetation Index (SVI) was Funding. This project was financed by the determined for a specific date throughout Peru, Instituto de Investigación para el Desarrollo as well as for two decades in three study points Sustentable de Ceja de Selva (INDES-CES) de la using the GEE. The change of the SVI values due Universidad Nacional Toribio Rodríguez de to causes such as El Niño Phenomenon for the Mendoza de Amazonas (Perú) mediante el Sechura Desert, Piura, and deforestation in Proyecto del Sistema Nacional de Inversión Tropical Forests in the La Pampa Zone, Madre de Pública (SNIP) N° 352431 “Creación de los Dios, was verified. Subsequently, in the Peruvian Servicios del Centro de Investigación en Andes, in Ollachea, Puno, it was determined that Climatología y Energías Alternativas” the most extreme negative SVI value represented (PROCICEA) y la Universidad Estatal de an extreme drought never recorded for this area. Montana (Estados Unidos). Finally, the GEE provided free of charge a 9 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Conflict of interests. The authors declare that Carbajal, C. M., Yarlequé, C., and Posadas, A., there is no conflict of interest related to this 2010. Datos faltantes de precipitación publication. pluvial diaria mediante la Transformada Wavelet. Revista Peruana Geo- Compliance with ethical standards. Nothing to Atmosférica, 88(2), pp. 76- declare/does not apply. 88. https://repositorio.senamhi.gob.pe/b itstream/handle/20.500.12542/1069/Re Data availability. Data is available from the construcci%C3%B3n-de-datos- corresponding author upon request. faltantes-de-precipitaci%C3%B3n- pluvial-diaria-mediante-la- Author contribution statement. Transformada- Jaris E. Veneros – Conceptualization, Data Wavelet.pdf?sequence=1&isAllowed= curation, Formal Analysis, Investigation, y Methodology, Software, Supervision , CENEPRED, 2020. Temporada de lluvia 2020 – Validation , Visualization , Writing – original 2021: escenario de riesgo por déficit de draft, Writing – review and editing., Ligia lluvias en la costa y sierra norte del Perú García – Project administration, Formal para el periodo setiembre – noviembre Analysis, Investigation, Methodology, 2020. pp. 1–22. En CENEPRED. Resources, Writing – original draft, Writing – Último acceso: 20 de diciembre, 2020 review and editing. de https://cenepred.gob.pe/web/wp- content/uploads/2020/09/ESCENARIO REFERENCES -DE-RIESGO-DEFICIT_PP.pdf Adedeji, O., Olusola, A., James, G., Shaba, H. Chuvieco, E., 2008. Earth observation of global A., Orimoloye, I. R., Singh, S. K., and change: the role of satellite remote Adelabu, S., 2020. Early warning sensing in monitoring the global systems development for agricultural environment. In Earth Observation of drought assessment in Nigeria. Global Change: The Role of Satellite Environmental Monitoring and Remote Sensing in Monitoring the Assessment, 192(12), pp. 1–21. Global Environment. Alcalá de https://doi.org/10.1007/s10661-020- Henares: Springer. 08730-3 COEN, 2020. Reporte de peligro inminente N° Anderson, E. P., Jenkins, C. N., Heilpern, S., 148 - 21/12/2020 / COEN - INDECI / Maldonado-Ocampo, J. A., Carvajal- 17:45 HORAS (Reporte N° 3) por Vallejos, F. M., Encalada, A. C., déficit hídrico en el departamento de Rivadeneira, J. F., Hidalgo, M., Cañas, Cajamarca. En INDECI. Último acceso: C. M., Ortega, H., Salcedo, N., 20 de diciembre, 2020 de Maldonado, M., and Tedesco, P. A., https://www.indeci.gob.pe/wp- 2018. Desierto de Sechura - Piura. content/uploads/2020/12/REPORTE- Science Advances, 4(1), pp. 1–8. DE-PELIGRO-INMINENTE- https://doi.org/10.1126/sciadv.aao1642 N%C2%BA-148-21DIC2020- PELIGRO-POR-DEFICIT-HIDRICO- Andina Agencia Peruana de Noticias, 2020. EN-EL-DEPARTAMENTO-DE- Comunidad de Ollachea tendrá CAJAMARCA-3.pdf participación de 5% en proyecto aurífero, señala Minera IRL. En Andina. Da Silva, C. A., Leonel-Junior, A. H. S., Rossi, Último Acceso: 25 de diciembre, 2020 F. S., Correia Filho, W. L. F., Santiago, de https://andina.pe/agencia/noticia- D. de B., de Oliveira-Júnior, J. F., comunidad-ollachea-tendra- Teodoro, P. E., Lima, M., and Capristo- participacion-5-proyecto-aurifero- Silva, G. F., 2020. Mapping soybean senala-minera-irl-370149.aspx planting area in midwest Brazil with remotely sensed images and phenology- Asurza, F., Ramos, C., and Lavado, W., 2018. based algorithm using the Google Earth Evaluación de los productos Tropical Engine platform. Computers and Rainfall Measuring Mission (TRMM) y Electronics in Agriculture, 169(2019), Global Precipitation Measurement pp. 1-10. (GPM) en el modelamiento hidrológico https://doi.org/10.1016/j.compag.2019. de la cuenca del río Huancané, 105194 Perú. Scientia Agropecuaria, 9(1), pp. 53- Didan, K., 2020. LP DAAC - MOD13Q1 62. https://doi.org/10.17268/sci.agrope MODIS/Terra vegetation indices 16- cu.2018.01.06 10 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Day L3 global 250 m SIN Grid. En Pinterest. Último acceso: 01 de USGS. Último acceso: 20 de diciembre, diciembre, 2020 de 2020 de https://www.pinterest.com/pin/5698461 https://lpdaac.usgs.gov/products/mod13 59084164271/ q1v006/ González-Moradas, M. del R., and Viveen, W., Didan, K., Munoz, A. B., Solano, R., and Huete, 2020. Evaluation of ASTER GDEM2, A., 2020. MODIS vegetation index user SRTMv3.0, ALOS AW3D30 and ’s guide (Collection 6). En Modis Active TanDEM-X DEMs for the Peruvian Fire and Burned Area Products. Último Andes against highly accurate GNSS acceso: 20 de diciembre, 2020 de ground control points and https://modis- geomorphological-hydrological fire.umd.edu/files/MODIS_C6_Fire_U metrics. Remote Sensing of ser_Guide_C.pdf Environment, 237(2020), pp. 1-19. https://doi.org/10.1016/j.rse.2019.111509 Ezzine, H., Bouziane, A., and Ouazar, D., 2014. Seasonal comparisons of Gorelick, N., Hancher, M., Dixon, M., meteorological and agricultural Ilyushchenko, S., Thau, D., and Moore, droughtindices in Morocco using open R., 2017. Google earth engine: short time-series data. International planetary-scale geospatial analysis for Journal of Applied Earth Observation everyone. Remote Sensing of and Geoinformation, 26(1), 36–48. Environment, 202, pp. 18–27. https://doi.org/10.1016/j.jag.2013.05.005 https://doi.org/10.1016/j.rse.2017.06.031 Fernández, C., 2020. Puno: ausencia de lluvias en Gutiérrez, C., Ochoa, L., and Velasco, I., 2005. el altiplano preocupa a los agricultores. Sequía, un problema de perspectiva y En El Comercio. Último acceso: 21 de gestión. Región y Sociedad, 17(34), pp. diciembre, 2020 de 35–71. https://elcomercio.pe/peru/puno/puno- http://www.scielo.org.mx/pdf/regsoc/v ausencia-de-lluvias-en-el-altiplano- 17n34/v17n34a2.pdf preocupa-a-los-agricultores-noticia/ INGEMMET, 2020. Geología del Perú. En GADM, 2020. Database of global administrative Instituto Geológico, Minero y areas version 2.8 (GADM). En GADM Metalúrgico - INGEMMET. Último maps and data. Último acceso: 21 de acceso: 01 de diciembre, 2020 de diciembre, 2020 de https://gadm.org/ https://repositorio.ingemmet.gob.pe/ha ndle/20.500.12544/176 Garcia, E., and Otto, M., 2015. Caracterización ecohidrológica de humedales alto Isminas Revista, 2020. La desgracia de vivir en andinos usando imágenes de satélite la riqueza. En La Oreja Roja. Último multitemporales en la cabecera de acceso: 01 de diciembre, 2020 de cuenca del Río Santa, Ancash, Perú. https://www.laorejaroja.com/la- Ecología Aplicada, 14(2), 2015. desgracia-de-vivir-en-la-riqueza/ http://www.scielo.org.pe/pdf/ecol/v14n Ji, L., and Peters, A. J., 2003. Assessing 2/a04v14n2.pdf vegetation response to drought in the García, L., Veneros, J., and Tineo, D., 2020. northern Great Plains using vegetation Severe acute respiratory syndrome and drought indices. Remote Sensing of (SARS-CoV-2): A national public Environment, 87(1), pp. 85–98. health emergency and its impact on food https://doi.org/10.1016/S0034- security in Peru. Scientia Agropecuaria, 4257(03)00174-3 11(2), pp. 241–245. Kibret, K. S., Marohn, C., and Cadisch, G., 2020. https://doi.org/10.17268/SCI.AGROPE Use of MODIS EVI to map crop CU.2020.02.12 phenology, identify cropping systems, Ghazaryan, G., Dubovyk, O., Graw, V., Kussul, detect land use change and drought risk N., and Schellberg, J., 2020. Local-scale in Ethiopia–an application of Google agricultural drought monitoring with Earth Engine. European Journal of satellite-based multi-sensor time-series. Remote Sensing, 53(1), pp. 176–191. GIScience and Remote Sensing, 57(5), https://doi.org/10.1080/22797254.2020. pp. 704–718. 1786466 https://doi.org/10.1080/15481603.2020. Lavado-Casimiro, W., and Espinoza, J. C., 2014. 1778332 Impactos de el niño y la niña en las GMC, 2020. Desierto de Sechura - Piura. En lluvias del Perú (1965-2007). Revista 11 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Brasileira de Meteorologia, 29(2), pp. intensa en América del Sur desde 2002. 171–182. En La República. Último acceso: 15 de https://doi.org/10.1590/S0102- diciembre, 2020 de 77862014000200003 https://larepublica.pe/ciencia/2020/11/0 9/esta-es-la-segunda-sequia-mas- MINAM, 2020. Mapa nacional de ecosistemas intensa-en-america-del-sur-desde-2002/ del Perú. En SINIA-Sistema Nacional de Información Ambiental. Último acceso: SENAMHI, 2020. a. Continuará la ausencia de 01 de diciembre, 2020 de lluvia en sierra central y sur. En https://sinia.minam.gob.pe/mapas/mapa Plataforma digital única del Estado -nacional-ecosistemas-peru Peruano. Último acceso: 13 de diciembre, 2020 de NASA, 2020. Severe drought in South America. https://www.gob.pe/institucion/senamhi/n En Earth Observatory. Último acceso: oticias/317152-continuara-la-ausencia-de- 01 de diciembre, 2020 de lluvia-en-sierra-central-y-sur https://earthobservatory.nasa.gov/imag es/147480/severe-drought-in-south- SENAMHI, 2020b. Mapa climático del Perú. En america SENAMHI. Último acceso: 13 de diciembre, 2020 de PCM, 2020. Decreto Supremo que declara el https://www.senamhi.gob.pe/servicios/ Estado de Emergencia en varios ?p=mapa-climatico-del-peru distritos de algunas provincias de los departamentos de Tumbes, Piura, SENAMHI, 2020c. Reporte climático de lluvias Lambayeque, La Libertad y Cajamarca, a nivel Nacional: setiembre – noviembre por peligro inminente ante déficit 2020. En Plataforma digital única del hídrico. En El Peruano. Último acceso: Estado Peruano. Último acceso: 13 de 15 de diciembre, 2020 de diciembre, 2020 de https://busquedas.elperuano.pe/normasl https://www.gob.pe/qu/institucion/sena egales/decreto-supremo-que-declara-el- mhi/informes-publicaciones/1412533- estado-de-emergencia-en-vario- reporte-de-climatico-de-lluvias-a-nivel- decreto-supremo-n-149-2020-pcm- nacional-setiembre-noviembre-2020 1884523-1/ Shamshirband, S., Hashemi, S., Salimi, H., Peters, A. J., Walter-Shea, E. A., Ji, L., Viña, A., Samadianfard, S., Asadi, E., Shadkani, Hayes, M., and Svoboda, M. D., 2002. S., Kargar, K., Mosavi, A., Nabipour, Drought monitoring with NDVI-based N., and Chau, K. W., 2020. Predicting standardized vegetation index. Standardized Streamflow index for Photogrammetric Engineering and hydrological drought using machine Remote Sensing, 68(1), pp. 71–75. learning models. Engineering Applications of Computational Fluid Phillips, O. L., Aragão, L. E. O. C., Lewis, S. L., Mechanics, 14(1), pp. 339–350. Fisher, J. B., Lloyd, J., López-González, https://doi.org/10.1080/19942060.2020. G., Malhi, Y., Monteagudo, A., 1715844 Peacock, J., Quesada, C. A., Van Der Heijden, G., Almeida, S., Amaral, I., Silva, C., Heinrich, V., Freire, A., Broggio, I., Arroyo, L., Aymard, G., Baker, T. R., Rosan, T., Doblas, J., Anderson, L., Bánki, O., Blanc, L., Bonal, D., and Rousseau, G., Shimabukuro, Y., Silva, Torres-Lezama, A., 2009. Drought C., House, J., and Aragão, L., 2020. sensitivity of the amazon rainforest. Benchmark maps of 33 years of Science, 323(5919), 1344–1347. secondary forest age for Brazil. https://doi.org/10.1126/science.1164033 Scientific Data, 7(1), pp. 1–9. https://doi.org/10.1038/s41597-020- Pulgar, J., 2014. Las ocho regiones naturales del 00600-4 Perú. Terra Brasilis, 3, 1–19. https://doi.org/10.4000/terrabrasilis.1027 Skakun, S., Kussul, N., Shelestov, A., and Kussul, O., 2016. The use of satellite Risco Sence, E., Casimiro, W. L., and Mejia data for agriculture drought risk Marcacuzco, A., 2017. Runoff quantification in Ukraine. Geomatics, generation at Sub-Basin scale in Peru Natural Hazards and Risk, 7(3), pp. (Pacific and Titicaca Basins). En E- 901–917. Proceedings of the 38th IAHR World https://doi.org/10.1080/19475705.2015. Congress. Ciudad de Panamá, Panamá., 1016555 1–6 Septiembre. pp. 4702-4710 Son, N. T., Chen, C. F., Chen, C. R., Minh, V. Rodell, M., 2020. Esta es la segunda sequía más 12 Tropical and Subtropical Agroecosystems 25 (2022): #027 Veneros and García, 2022 Q., and Trung, N. H., 2014. A Disaster Management and Emergency comparative analysis of multitemporal Response (UN-SPIDER). Último MODIS EVI and NDVI data for large- acceso: 13 de diciembre, 2020 de scale rice yield estimation. Agricultural https://un-spider.org/advisory- and Forest Meteorology, 197, pp. 52– support/recommended- 64. practices/recommended-practice- https://doi.org/10.1016/j.agrformet.201 agricultural-drought-monitoring-svi 4.06.007 Wainwright, H. M., Steefel, C., Trutner, S. D., Spinoni, J., Barbosa, P., De Jager, A., Henderson, A. N., Nikolopoulos, E. I., McCormick, N., Naumann, G., Vogt, J. Wilmer, C. F., Chadwick, K. D., Falco, V., Magni, D., Masante, D., and N., Schaettle, K. B., Brown, J. B., Mazzeschi, M., 2019. A new global Steltzer, H., Williams, K. H., Hubbard, database of meteorological drought S. S., and Enquist, B. J., 2020. Satellite- events from 1951 to 2016. Journal of derived foresummer drought sensitivity Hydrology: Regional Studies, 22(2019), of plant productivity in Rocky Mountain pp. 1-24. headwater catchments: spatial https://doi.org/10.1016/j.ejrh.2019.100593 heterogeneity and geological- geomorphological control. Swain, S., Wardlow, B., Narumalani, S., Environmental Research Letters, 15(8), Tadesse, T., and Callahan, K., 2011. pp. 1–10. https://doi.org/10.1088/1748- Assessment of vegetation response to 9326/ab8fd0 drought in nebraska using Terra- MODIS land surface temperature and Wang, D., Morton, D., Masek, J., Wu, A., Nagol, normalized difference vegetation index. J., Xiong, X., Levy, R., Vermote, E., GIScience and Remote Sensing, 48(3), and Wolfe, R., 2012. Impact of sensor pp. 432–455. degradation on the MODIS NDVI time https://doi.org/10.2747/1548- series. Remote Sensing of Environment, 1603.48.3.432 119, pp. 55–61. https://doi.org/10.1016/j.rse.2011.12.001 Tarnavsky, E., and Bonifacio, R., 2020. Drought risk management using satellite-based Xiangtao, W., Xianzhou, Z., Junhao, W., and rainfall estimates. In Advances in Ben, N., 2020. Variations in the drought Global Change Research. Reading: severity index in response to climate Springer. change on the Tibetan Plateau. Journal of Resources and Ecology, 11(3), pp. Tsakiris, G., and Vangelis, H., 2004. Towards a 304–314. Drought Watch System based on spatial https://doi.org/10.5814/j.issn.1674- SPI. Water Resources Management, 764x.2020.03.008 18(1), pp. 1–12. https://doi.org/10.1023/B:WARM.0000 Zargar, A., Sadiq, R., Naser, B., and Khan, F. I., 015410.47014.a4 2011. A review of drought indices. Environmental Reviews, 19(1), 333– UN-SPIDER 2020. Drought monitoring using 349. https://doi.org/10.1139/a11-013 the Standard Vegetation Index (SVI). En Space-Based Information for 13