Estimation of the Indonesian drought based on phenology vegetation analysis of maize

Climate change has a widespread influence on food essentials. Climate change is a pressing issue in Indonesia, a rapidly developing country. As a result, to forecast dryness during maize production in Indonesia’s Central East Java districts, it was proposed that a study be done on vegetation phenology. The study uses the vegetation to calculate the NDWI (Water Index) and Temperature (LST), calculate drought severity based on precipitation data. NDWI and LST were obtained from Landsat 8 OLI with a spatial resolution of 30 m and were used to identify water shortage and temperature within the study area. The inquiry was finished in 2018 and showed that the vegetation phenology was based on the growing season. The growing season of maize report from the NDVI (Vegetation Index) trajectories that April 2018 was the maize planting season, and the harvest was completed in later August. Additionally, LST analysis found that the temperature was higher in mid and southern Central East Java, Indonesia. To validate the data, rainfall information was used to compute the drought severity using SPI method to identify drought-prone areas. Drought severity validation data were validated for vegetation phenology analysis in 2018.


Introduction
In recent decades, there has been an increasing body of evidence demonstrating that Climate change has several direct and indirect implications on global and individual health [1].Climate has a large impact on population dynamics and migratory patterns is significant.It also influences population growth, development, and dispersal trends.Temperature and precipitation changes may have an influence on plant phenology, the severity of winters, drought and wildfire conditions, the distribution and abundance of invasive species, predation, and disease in a direct or indirect manner.Indirect implications of these changes include the spread and abundance of exotic species, predation, and sickness [2].As a result of global warming, sea levels may rise, and rainfall patterns may vary.Furthermore, it may increase the frequency of extreme weather events such as floods, forest fires, and droughts., which can disrupt public health infrastructure and increase mortality, morbidity, and disability, as well as the incidence of infectious diseases and psychiatric disorders [3].Due to the variability of rainfall and the scarcity of data, obtaining comprehensive and reliable rainfall observation data is challenging [4][5][6].One of them used GSMaP (Global Satellite Mapping of Precipitation) satellite data with high spatial and temporal resolution to tackle this problem, demonstrating that satellite technology can collect rainfall data for a region area [7].However, because satellite 1230 (2023) 012144 IOP Publishing doi:10.1088/1755-1315/1230/1/012144 2 measurements are indirect, the data contains more uncertainties because cloud top reflectivity, heat radiation, as well as intermittent satellite communications overflights are all inaccurate [8].
To characterize patterns of change in vegetation phenology, two factors can be used: phenology, which is used as a landscape predictor of variability in plant response over time, and basic environmental variables, which consistently drive vegetation dynamics in a specific location year after year [9].Drought has also harmed one of Central East Java's most important maize harvests [10].Droughts can be caused by weather, agriculture, hydrology, or economic factors [11,12].There are presently no services in Indonesia that give detailed information about the drought to the public.The Indonesian Meteorology, Climatology, and Geophysics Agency provides the sole meteorological drought warnings (BMKG).A meteorological drought is defined as a temporary decrease in precipitation relative to the average [13].In meteorological, drought that occurs during sub-normal precipitation in an area during a certain period, agricultural drought generally a period of reduced soil moisture and resultant crop failure due to lack of surface water.Agricultural drought definitions also relate different characteristics of weather drought.Hydrological drought is associated with the effects of reducing the flow frequency of surface or ground water (such as rivers, reservoirs, lakes, and groundwater) [14].The drought's severity is defined by the ratio of precipitation to what is considered typical for that time, and the length is calculated by how long the dry state lasts.A meteorological drought has an impact on agriculture by limiting the quantity of dam water available for irrigation and causing a groundwater shortage, which may lead to crop failure.If the meteorological drought continues, a hydrological drought, defined as a decreased supply of surface water and groundwater based on river, reservoir, lake, and groundwater levels, might occur.
Central East Java was the most impact in the drought-prone areas based on the weather monitoring system.The monitoring for drought areas were utilized the vegetation phenology with Google Earth Engine (GEE) [15].GEE is a cloud platform that enables global geospatial data parallel processing [16] and provides remote sensing for massive data analysis [17].GEE stands for geospatial processing service.Earth Engine can do large-scale geospatial processing that includes geophysical, climate and weather, and imagery facilities, as well as ready products like Terrain, Surface Temperature, Atmospheric, Cropland, Land Cover and other satellite image sensors [18].GEE is a web-based code editor for data exploration, visualization, and categorization on cloud platform [19,20].
The BMKG agency usually forecast the drought by the rainfall event.In this study we purpose to determine which data sets and algorithms are more efficient and reliable in time-series analysis of vegetation phenology and predicting weather for drought periods during the maize growing season and validated the vegetation with the assessment of the precipitation data.

Study area
Central East Java, which includes Gresik, Lamongan, Malang, Mojokerto, Pasuruan, Sidoarjo, Tuban, Batu City, Malang City, Mojokerto City, Pasuruan City, and Surabaya City, is made up of twelve regencies.Rice, maize, cassava, and a variety of vegetables are all important crops.Irrigation water was supplied by rivers flowing from the nation's northern region, which is also crossed by the Surabaya River to the south.

Methods
The image was evaluated and validated using satellite remote sensing in the study.The satellite image was created using ArcGIS 10.4.1® and incorporates the vegetation phenology of the NDVI, NDWI, and the land surface temperature (LST) from Landsat 8 OLI data.To guarantee data accuracy, we compare two sets of precipitation data from two different sources: the Local Station and Global Rainfall MAP (GSMap, JAXA) Agency,BMKG.The use of data and precipitation this year was necessitated by a drought in 2018.Figure 1 shows an example.

Vegetation Phenology.
Phenology is a valuable approach for studying the cycles of vegetative development.Phenology is sensitive to climate change and provides important information for researching trends in biological processes or climatology, which may track the impacts of climate change at various world scales.Landsat 8 OLI was used to compute vegetation phenology such as NDVI, NDWI, and LST.Maize phenology is an important topic for farmers since it affects the time of planting, fertilization, and harvesting and it is a critical phase that varies based on maize type, climate, and local growth circumstances.Farmers can use a variety of strategies to keep track of the phenology of their maize crops.They can, for example, use satellite imagery to examine plant emergence and development, or to estimate pollination time by measuring the growth of tassels and silks.Using this information, they could plan crop planting and harvesting, as well as fertilizer and other input applications.

Normalized Difference Vegetation Index (NDVI).
The NDVI may be used to depict horizontal and vertical geographical and temporal distributions, as well as vegetative cultivation and biomass, and a few vegetational properties, such as plant features, production, and temperature impacts on the plant.GEE was utilized for high-performance computation to investigate vegetation phenology in equation 1. Figure 2 shows an example.NDVI = (NIR -Red) / (NIR + Red) (1) where: NIR stands for near-infrared reflectance and Red represents the red reflectance value.
where: Green stands for green reflectance and SWIR stand for short-wave infrared reflectance.

LST.
The LST is an element in formulating regional energy and water budgets, several environmental factors, including biodiversity, drought, urban growth, and climate change.

Data access of precipitation.
Global Rainfall MAP (GSMap, JAXA) and BMKG Agency are the two sources of precipitation dataset.The GSMAP were obtained from the website https://sharaku.eorc.jaxa.jp/GSMaP/and the local station were obtain from BMKG Agency.

Global precipitation datasets.
Because it is based on a microwave radiometer algorithm that is compatible with a precipitation radar algorithm, GSMaP's overall precipitation map has a high temporal and spatial resolution (Figure 3).GSMaP, which is funded by JAXA and JST.GSMaP has created a worldwide precipitation map based from many satellites as part of the Global Precipitation Measurement (GPM) Mission, which uses dual-frequency precipitation radar (DPR) within the GPM core spacecraft.The datasets were updates hourly with global precipitation statistics.

Local precipitation datasets.
The data precipitation gauge from the precipitation analysis were provided by the BMKG Agency.Precipitation observation data for the study area's stations, including the Tuban Regency station, were available.These stations were set up using 2016 rain gauge data, and the rain gauge was useful for analyzing precipitation monitoring in Figure 4.The Tuban Regency experienced a dry year in 2018, with February receiving the most rain and may receiving the least.
During the reference period, the average trend in Tuban Regency was 13 mm per month.The most precipitation fell in the first month of 2018.Precipitation estimates for the study region indicate a drop between April and September and an increase between October and February.The total amount of precipitation was compared to worldwide precipitation databases.

Vegetation indices analysis
Vegetation indicators are used to measure maize status during the growing season.The 2018 NDVI agricultural season runs from April until the end of August.According to additional indicators such as the NDWI, which was found in April 2018 (Figure 5), the growing season and maize planting were both dry.LST was enhanced in a same manner in the southern and central regions of Indonesia's East Java (Figure 6).
The World Meteorological Organization (WMO) claims that the 3-month SPI is more useful for emphasizing the presence of moisture in agricultural regions [21] in Table 1.SPI compares precipitation over three months to the entire year, which is part of the long-term historical data.Moderately moisture -0.99 -0.99About average -1.0 -1.49 Moderately arid -1.5 -1.99 Severely arid -2.0 and less Extremely arid In this study, we compared vegetal phenology in drought-prone regions to a three-month SPI.According to SPI data, the months of April through October of 2018 were dry.The SPI fluctuated between -1.21 to -1.34 throughout this time period, as seen in table 1.Following a drop in September, the SPI index fell even further in October (See Figure 7).

Conclusion
This study was conducted at the provincial level to categorize areas that are vulnerable to drought for maize production using vegetation phenology and a drought severity index.The analysis of vegetation indices was required for processing the vegetation phenology.The NDVI, NDWI, and LST were employed in the phenology.The NDVI can be used to determine both the vegetation profile pattern and the growing season of a plant.NDWI was used to assess the region's water scarcity for agricultural reasons, while LST was utilized to correlate it with temperature.The severity index was generated using SPI, whereas the phenology of the plants was determined using NDWI and LST.According to the findings, the comparison of vegetation phenology and drought severity took place concurrently between April and October of this year, which is the maize growing season.In vegetation analysis, we saw for NDVI could identify the growing season of crop where in the study we focus for maize crops, and NDWI could identify the waterbody.In figure 5, the reproductive stage in July 2018 the waterbody shows red not blue in the southern area or in Malang District.In figure 6, LST helps to understand the temperature surface of the land where in Malang and in the middle Central East Java show dark chocolate color.The validation data were compared using Jaxa Gauge and Local Station, the results on August show SPI with -1.77 among all part of this region in Central East Java.With high temperatures and little rain, stakeholders and farmers can use this research to help prevent a drought.
A more thorough analysis using vegetation phenology [22] and object detection [23,24] will be used to validate the yield model.Predicting vegetation phenology and drought severity also assists in the prevention of crop decreases and the provision of local food security.

Figure 1 .
Figure 1.Drought forecasting framework for Central East Java.

Figure 4 .
Figure 4.The Selection of Local Precipitation in Central East Java.

Figure 5 .
Figure 5. From left to right, monitoring growing season Maize analyze utilizing NDWI in Central East Java (a) Planting, (b) Reproductive, (c) Harvest.

Figure 6 .
Figure 6.From left to right, monitoring growing season Maize analyze utilizing LST in Central East Java during the growing season (a) Planting, (b) Reproductive, (c) Harvest.

3. 2 .
Validation data We use GSMap and Local Station to validate the data.The Indonesia Meteorological and Climate Agency provided local stations, and JAXA in Japan developed GSMap.The Standard Precipitation Index was used to calculate both precipitations (SPI).The gamma function, as seen below, represented the SPI in the cumulative distribution (3): (  ) = ∫ (  )