Introduction

India encompasses a wide range of ecosystems: desert, high mountains, highlands, tropical and temperate forests, swamplands, plains, grasslands, wetlands, mangroves, coral reefs, and island archipelagos. There are four biodiversity hotspots in the country: the Western Ghats, the Himalayas, the Indo-Burma region, and the Nicobar Islands. India harbors 7–8% of globally recorded species in 2.4% of the land area; many species are endemic. Biodiversity underpins all ecosystem goods and services responsible for providing food, water, and medicines, buffering the impacts of climate change, controlling the outbreak of diseases, and supporting nutrient cycling (Xie et al. 2017; Babbar et al. 2020) (Supplementary Fig. 1). Chen and Sun (2018) also explained the importance of botanical gardens as the sites for scientific research, in/ex-situ conservation, genetic studies and propagation of seeds, taxonomic and systematic studies of the present species in the garden, environmental restoration, public awareness, and guidance.

The botanical gardens have a role in the exploration of plant diversity. Bringing these parameters under the concept of a botanical garden has taken many centuries. During the twenty-first century, the focus included not only the conservation of rare and threatened plants and their seed for germplasm collection, but also climate change, food security, public awareness, and sustainability also came into play. (Borsch and Lohne 2014; Krishnan and Novy 2016; Gaio-Oliveira et al. 2017; Areendran et al. 2020).

A few names of the global botanic gardens include Royal Botanic Garden (Kew), Chicago Botanic Garden, USA, Brooklyn Botanic Garden, USA, Singapore Botanic Garden, Singapore, and National Botanic Garden (Lucknow, India), and Lloyd Botanic Garden (Darjeeling, India) are some of the names in India. According to the Botanical Garden Conservation International (BGCI), the total number of accessions recorded in botanical gardens of the world is ca. 6 million, out of which India has approximately 2,00,000 (Chen and Sun 2018). There are 122 botanical gardens spread all across India, encompassing around 4% of the global botanical gardens (Golding et al. 2010).

Out of these 122 gardens, one garden named Botanic Garden of Indian Republic (BGIR), is situated in Noida, Uttar Pradesh. It has been functional since 2002; it comes under the Ministry of Environment Forest and Climate Change (MoEF&CC) and is maintained by the Botanical Survey of India. The garden is surrounded by high human density, and hence it faces the effects of anthropogenic activities such as transportation dust and vehicular and industrial emissions that cause pollution.

The anthropogenic activities have an impact on biomass production, and plant and soil health. In order to understand these environmental changes, the spatial trend of soil nutrients is essential in understanding the terrestrial ecosystem and the temporal changes associated with it (Liu et al. 2021; Mallick et al. 2022; Calzolari et al. 2020). Kren (2018) revealed that the physical and chemical properties of the soil are determined by the spectral reflectance of the soil. Satellite image consists of different bands based on wavelength. Here we use bands like visible (VIS 0.4–0.7 μm), near-infrared (NIR 0.7–1.1 μm), short wave infrared (SWIR 1.1–2.5 μm), and these wavelengths get reflected by the soil that helps in determining the soil characteristics like soil texture, pH, electrical conductivity (EC), soil composition i.e., organic carbon (OC), nitrogen (N), potassium (K), phosphorus (P), calcium (Ca), Magnesium (Mg), zinc (Zn), sulphur (S), manganese (Mn), boron (B), iron (Fe) and copper (Cu).

Xu et al. (2018a, b) determined the spatial pattern of total nitrogen in the soil using remote sensing and geographic information systems (GIS). The spectral reflectance of these electromagnetic radiations is a tool that is now replacing the wet soil chemistry, and it can be used when the vegetation cover is low so that the soil is exposed to these radiations (Hill et al. 2010; Soriano et al. 2014; Nocita et al. 2015; Escribano et al. 2017; Kren 2018). Soil chromophores help determine the presence of a particular element in the soil (Ben-Dor 2011; Escribano et al. 2017).Various studies used multispectral data from Landsat 8 (OLI) and sentinel 2A (MSI) to generate indices that help determine soil properties (Roy et al. 2014; Castaldi et al. 2015; Kren 2018).

The supporting ecosystem service provided by the soil is to reinforce biomass production by nutrient cycling (Wilcke et al. 2013; Mallick et al. 2022). When the soil capacity for producing biomass changes, it implies a change in soil quality, in other words, soil fertility determines the soil quality (Nguemezi et al. 2020). In order to check the soil fertility, we used different types of vegetation indices like Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI), Renormalized Difference Vegetation Index (RDVI), Normalized Difference Water Index (NDWI), Difference Vegetation Index (DVI), Normalized Difference Moisture Index (NDMI) and Normalized Pigment Chlorophyll Index (NPCRI) (Leon et al. 2003; Ali et al. 2016; Kaplan et al. 2017; Zhang et al. 2019). Different studies used various spatial indices to understand the different parameters and effects on the ground like urban heat island effect (Sarif et al. 2020; Pathak et al. 2021). Sarif and Gupta (2019) used indices like NDVI, Normal Difference Build-up index (NDBI), Enhanced Build-up and Bareness index (EBBI) to predict the land surface temperature (LST) NDVI detects the vegetation-covered areas and its value determines chlorophyll content and therefore a predictor of photosynthesis (Meeragandhi et al. 2015). This is a widely used index for determining vegetation status and soil nutrients in many areas using remote sensing (Bernardi et al. 2017). This index is sensitive to some soil properties present in the background of the vegetation i.e., soil color, soil brightness, surface moisture content, shade of vegetation canopy, and sometimes atmospheric conditions like a cloud. In order to avoid such circumstances, we use different other indices and make a comparative result for better accuracy (Masroor et al. 2022). The various indices which are sensitive to the soil line can be a good predictor of soil conditions and soil nutrient levels. The advanced vegetation indices like Atmospherically Resistant Vegetation Index (ARVI), Enhanced Vegetation Index (EVI), Moisture Adjusted Vegetation Index (MAVI) Modified Soil Adjusted Vegetation Index (MSAVI), are designed to reduce the soil biophysical effects and are preferred over NDVI. Other indices like Bare soil index (BSI), soil adjusted total vegetation index (SATVI) are also sensitive to soil or atmospheric conditions and therefore can be used for soil parametric study (Zhu et al. 2014; Li and Chen 2014; Wilson et al. 2016; Xue and Su 2017; Kaplan et al. 2017; Chakrabortty et al. 2020).

The idea behind the botanical garden is to do ex-situ conservation for which the soil of the garden plays an important role in the successful establishment of the plant (Faraji and Karimi 2020). Therefore, the objectives of our study are to check the present time soil status of BGIR and also suggest ways for its improvement for the future successful establishment of the introduced species; to find the correlation among the soil nutrients to study the dependency of the presence of one nutrient over the other and develop the mathematical equations using MLRA, in order to avoid the wet chemistry in future; to perform the Correspondence Analysis to study the association of the elements with the soil type of the study area and to find the correlation between the soil field data and the spatial data using statistical and remote sensing approaches; to understand the presence of any particular element in the soil based on the spectral reflectance of the soil.

Materials and methods

Study area

Botanic Garden of Indian Republic (BGIR) is situated between 26° '28' to 29° '30' N latitudes and 91° '30' to 97° '30' E longitude in the Gautam Budh Nagar district of Noida, Uttar Pradesh, India (Fig. 1). BGIR is surrounded by high human activity and hence it faces the effects of activities like high air pollution which leads to acid rain that ultimately affects the vegetation as well as the soil conditions of the area. Before the establishment of the garden, the whole area was cultivated with different crops.

Fig. 1
figure 1

Location map of the botanic garden of the Indian Republic (BGIR), Noida, Uttar Pradesh, India

Full size image

The area of the park is ca. 164 acres and the objective of the garden is to conserve the endangered and the introductory species. The botanic garden consists of eight man-made zones along with a medicinal plant section, taxonomic garden, rose garden, cactus section, and seed bank (Fig. 2, Tables 1, 2). Many endangered species of plant are well conserved in BGIR, such as Pterocarpus santalinus (Red Sandalwood) Hildegardia populifolia, Chukrasia tabularis (Mahogony) (Table 1). The complete list (Table 1) is derived from the past records of the BGIR. Table 2 consists of the plant species present in each zone and the identification and authentication of the plant species have been done by the taxonomists of the Botanical Survey of India (BSI) based on morphological studies of the plant. Local flora study and herbarium study have also been done for the verification of the species. Some of the alien herbaceous species like Parthenium hysterophorus, Lantana camara, etc. are occurring naturally in this garden along with other natural vegetation. Presently the garden has ca. 350 species (introduced) and ca. 100 wild species, which include herbs, shrubs, and trees including medicinal plants introduced from different parts of India. There are over 5000 trees belonging to 253 species, of which family Fabaceae constitute the maximum number of trees in woodland and other arboreta.

Fig. 2
figure 2

Different sections (fruit section, Charak Udyan section, economic plant section (EPS), medicinal section) and zones (zone 1–zone 8) of BGIR, Noida, Uttar Pradesh, India

Full size image
Table 1 Detailed list of ex situ conserved endangered plants along with the names of the family they belong (source: BGIR, Botanical Survey of India)
Full size table
Table 2 Detailed list of composition of plants in different zones (i.e. Zone 1–Zone 8 and other sections like Economical garden, Fruit Section, Charak Udhyan, and Cactus section) of BGIR
Full size table

This garden is located at the flood plain of river Yamuna, having an alluvial type of soil known as khaddar, which represents fairly coarse sand with a very fine texture on surface soil, greyish-brownish to grey depending upon depth. The climate of BGIR is a sub-tropical type, with three distinct summers, monsoon, and winter seasons. Noida experiences extreme climatic conditions, from a scorching dry summer (April–June) to a chilling winter (November–January). The monsoon is short (July–September), followed by the post-monsoon (September–October) season. The average temperature of the area (January 2002–December 2019) approaches a minimum of 20.2 °C and a maximum of 31.8 °C. The average annual rainfall recorded is 578.7 mm (Safdarjung weather station, New Delhi).

Soil management practices, sampling, and laboratory analysis

Given the period of establishment, this garden may have gone through more than 17 rainy and summer seasons during which wild vegetation may get dried, fall on the earth's surface, and enrich soil year after year by transforming the same back into the soil. The process occurs in areas which are left undisturbed to maintain natural ecosystem and in case of managed areas the nutrients taken out in the form of dead fallen leaves or wood are returned in the soil with the use of the vermicomposting unit. With this approach, we try to maintain the loss of nutrients from the soil.

Soil samples have been collected using stratified sampling techniques though it is not randomly stratified as we had to cover the different zones of the garden (Meng 2013). The soil samples of the garden were taken by Soil Auger from selected stations and sites between 09:00 AM and 12:00 Noon. Soils from different zones were kept separately to get zone-wise soil nutrient information. The samples were collected from 10 cm depth after the removal of surface leaf litter and kept in black poly bags labeled as per sites and sampling dates. Further, soil samples were brought to the laboratory, freeze-dried, and analyzed. The total sample collected was 25 from five different zones, i.e., five samples from each zone, and these samples were subjected to get the macro and micronutrient profile of this garden. In any ecosystem like Botanic gardens, the soil has a vital role in the growth of artificial plantations. The soil preparation and analysis were carried out by Indian Agricultural Research Institute (IARI), PUSA, New Delhi, Delhi. The methods used to do the soil analysis are given in Table 3. Carbon stock per unit area was calculated using the formula by Benbi and Brar (1992).

Table 3 The table represents the physico-chemical methods of extraction of the elements for the analysis of the soil parameters
Full size table
$$C\,{stock} (t/ha)=Carbon\,content\,(\%)\,\times\,Bulk \,density\,(Mg/{M}^{3})\,\times\,Depth\,(m)\,\times\,100$$
(1)

Satellite data extraction for the preparation of indices

The Landsat OLI image of 30 m resolution (21st Dec 2018) and Sentinel 2A image of 10 m resolution (9th Feb 2019) were taken for developing the indices from USGS. The data for both satellites were taken for post-monsoon season to avoid a mismatch in the results. Pre-processing of the satellite image consists of the atmospheric correction, which has been done to get the actual surface reflectance using the Fast line-of-Site Atmospheric Analysis of Spectral Hypercubes (FLAASH) tool in the ENVI 5.0 software (Zhang et al. 2020). Layer stacking has been performed on the downloaded image. The area of interest, i.e., BGIR, has been taken out from the image using the shapefile. Different formulas (Table 4) have been used to prepare the 'index's maps. The mean value for the respective zones has been used for the statistical analysis. These indices values have been used for understanding the relationship between the remotely sensed data and the field soil nutrient data based on the reflectance of the soil in a different wavelength of light.

Table 4 Detailed listing of the formula used for 14 different kinds of spatial indices for the analysis of the soil
Full size table

Statistical design and analysis

Based on descriptive statistics and normality test, the electrical conductivity, phosphorus, potassium, sulphur, and zinc were log-transformed to follow a normal distribution, and the other variables were kept the same. Pearson's correlation bivariate analysis has been performed to find if there is any correlation among the variables at a 95% confidence interval (P < 0.05) using R software 4.0.2(Ratnayake et al. 2019). The package used for finding the correlation was “corrplot” Multilinear regression analysis (MLRA) to investigate the relationships among the correlated variables. We have chosen MLRA over simple linear regression analysis because of the presence of more than one independent variable. The low P-value tells us about the significance of the test, and the high R2 value explains about variation between the variables (Patle et al. 2018).

The prediction model for MLRA analysis is based on the dependent and independent variables along with the regression coefficients (Egbe et al. 2017). In our study, soil pH, organic carbon (OC), and electrical conductivity (EC) have been used as the independent variables, and others are used as the dependent variables. Model validation has been done using root mean square error (RMSE). The difference between the observed value (y) and the predicted value (ý) has been considered for finding out the RMSE value. RMSE tells about the distribution of the data around the fitted regression line. RMSE determines the residual standard deviation (Patle et al. 2018).

Correspondence Analysis (CA) has been performed using Software XLSTAT 2020 by Addinsoft. CA tells about the association between the categorical variables by allowing visual interpretation in a low-dimensional space (Sourial et al. 2010; Greenacre 2015). The derived graph gives an overall picture of relationships among row and column pairs, which cannot be understood from the pairwise analysis. The nominal variable in our study are zones of BGIR (rows) from where the soil sample has been taken and the physicochemical parameters analysed. In the second phase of the study, 'Pearson's correlation coefficient using package "corrplot" of R software 4.0.2 was used to find the level of association among all the spatial indices and the field soil data at P < 0.05 level of significance.

Predictor variables

Predictor variables are the independent variables on which the response variables depend. The independent variables in this study are pH, EC, carbon %, and keeping potassium (Schroth et al. 2003; Nishanth and Biswas 2008), nitrogen (Mulder et al. 2002), iron (Boukhalfa and Crumbliss 2002; Colombo et al. 2013), boron (Yoshinaga et al. 2001), zinc (Alloway 2008), manganese (Nord and Lynch 2009), sulphur (Probert 1980; Gupta and Germida 1989; Binkley et al. 1989; Serrano et al. 1999) as dependent variables.

The indices with no effect of background soil conditions and others having significant effects of soil background conditions have been chosen to understand the relationship between vegetation indices and soil nutrients. All the indices have different adjustment factors, like soil adjustment, and moisture adjustment for better predictor variables of soil status. Spatial predictor variables are NDVI, EVI2, MAVI, MSAVI, SATVI, BSI, SCI, NDWI, DVI, ARVI, NDMI, NPCRI, and RDVI. These spatial indices are predictor variables for the soil physicochemical properties of the soil of BGIR (Xu et al. 2018a, b).

Results

Descriptive statistics and Pearson's correlation statistics among the soil elemental data

Descriptive statistics of any data bring the data into an easily understandable and manageable form as it gives the measure of central tendency of the data. The mean value for each soil physico-chemical parameter along with its SE has been mentioned. The soil pH value ranges from 8.07 to 8.31 (mean = 8.2 ± 0.5), EC ranges from 0.09 to 0.29 (mean = 0.15 ± 0.04), OC ranges from 0.16–0.38 (mean = 0.24 ± 0.04), N in the soil ranges from 105–168 (mean = 126.8 ± 12.2), P has a range of 6.7–13.4 (mean = 8.95 ± 1.19), K has a range from 83.7–449 (mean = 189.68 ± 67.29), Ca range lies from 1976–2362 (mean = 2165.6 ± 77.2), Mg in the soil has a range of 89.3–218 (mean = 163.86 ± 21.73), S ranges from 8.19–61.8 (mean = 22.5 ± 10.2), Zn has a range of 0.26–9.56 (mean = 2.62 ± 1.74), Fe in the soil has a range of 2.6–3.92 (mean = 3.3 ± 0.2), Cu ranges from 0.64–1.24 (mean = 0.98 ± 0.1), Mn range lies from 2–2.72 (mean = 2.34 ± 0.12), boron range is 0.22–0.47 (mean = 0.37 ± 0.04) in the soil. Other useful statistics like SD, SE, and skewness can be seen in Table 5.

Table 5 Descriptive statistics of the soil elements, showing number of samples taken, range, maximum and minimum value, mean, standard deviation, variance, and skewness
Full size table

The Pearson correlation coefficient at P < 0.05 is considered significant for further analysis, taking into account the predictor variables, i.e. OC (%), EC, and pH. In Fig. 3 the nutrients found correlated are S and EC (r = 0.97), N and OC (r = 0.98), Mn and OC (r = 0.97), Mn and N (r = 0.92), Ca and pH (r =  − 0.91), Cu and Fe (r = 0.94).

Fig. 3
figure 3

Pearson correlation coefficient of the soil data at P < 0.05 (red line to blue line value ranges from  − 1 to + 1 (i.e., − 1, − 0.8, − 0.6, − 0.4, − 0.2, 0, 0.2, 0.4, 0.6, 0.8, 1 from red line to blue line)). The circles represent robust relationship between a nitrogen v/s organic carbon (r = 0.98), b Log10 sulphur v/s Log10 electrical conductivity (r = 0.97), c calcium v/s pH (r = 0.91), d manganese v/s nitrogen (r = 0.92), e manganese v/s organic carbon (r = 0.97), f iron (Fe) v/s copper (Cu) (r = 0.94)

Full size image

Regression prediction modeling

Equations derived from MLRA are the reference for future studies in BGIR. Adjusted R2 is an actual predictor of the quality of the regression model as R2 does not consider the number of variables, which gives us the degree of determination. The various nutrients show dependency on the independent variables, as shown in Fig. 4. The results of the prediction model are as follows (Table 6):

$${\text{Log}}_{10} S = 1.7778*{\text{Log}}_{10} \;{\text{EC}} + 2.7258\quad (P < 0.05,\;R^{2} = 0.91)$$
(2)
$$N = 297.79*{\text{OC}}\left( \% \right) + 54.14\;(P < 0.05,\;R^{2} = 0.96)$$
(3)
$${\text{Mn}} = 2.8905*{\text{OC}}\left( \% \right) + 1.6347\;(P < 0.05,\;R^{2} = 0.94)$$
(4)
$${\text{Ca}} = - 1447.6*{\text{pH}} + 14044\quad (P < 0.05,\;R^{2} = 0.84)$$
(5)
Fig. 4
figure 4

Graphs representing relationship between a nitrogen v/s organic carbon, b manganese v/s organic carbon, c sulphur v/s electrical conductivity, d calcium v/s pH

Full size image
Table 6 Results of regression prediction model of various nutrients
Full size table

Correlations based on elements other than pH, EC, and OC (%) predictor variables have not been considered for the regression analyses because of the literature deficiency, supporting the fact that N, Cu, and Fe have been used as the predictor variables.

Validation of the results

The results have been validated with the high R2 and low RMSE values. The variables like N and Mn are positively correlated with OC at RMSE 4.5, 0.14. Sulphur has shown positive relation with EC at RMSE 0.35, while pH has shown a negative correlation with Ca at P < 0.05 significance level and RMSE 62.73 (Table 7).

Table 7 Validation statistics of the regression model based on the RMSE values
Full size table

Correspondence analysis (CA) of soil parameters concerning different zones of BGIR

Correspondence analysis is used to statistically analyze and graphically display the relationships among different sites (rows) and soil parameters (columns). It is calculated based on the factorization of the total parameters' values involved in the study. Also, it depends on the Eigenvalues plotted against the factors pair. The CA shows (Fig. 5) that Zone 3 has a high association with the EC and S and Zn has an association with Zone 5 and the taxonomy garden zone, and the rest of the parameters are clustered around the origin, hence showing less association concerning the study sites. The results are reliable, showing a 93.81% variance in the data.

Fig. 5
figure 5

Correspondence analysis between soil elements (column) and study site (row)

Full size image

Soil parameters and spatial indices correlation

The various spatial indices are shown in Fig. 6a–c, and their correlation with the soil parameters is shown in Fig. 7. The NDVI, EVI2, ARVI, NPCRI, RDVI have shown significant correlation with the Log10S, Mg, Log10Zn, B, and Fe respectively (with respective Pearson correlation coefficient r = 0.88, r =  − 0.90, r =  − 0.93, r = 0.91, r = 0.92 at P < 0.05) while along with ARVI the indices SCI, NDMI, MSAVI are also the predictor variable for Log10Zn (r =  − 0.89, r =  − 0.88, r = 0.92 at P < 0.05 respectively). MAVI2 positively correlate with OC, Mn content (r = 0.91, r = 0.93 respectively). MSAVI is negatively interrelated with Ca (r =  − 0.89) and SCI is negatively interrelated with Log10 K (r =  − 0.98), BSI is positively associated with pH (r = 0.91) and negatively with Ca (r =  − 0.93), while other indices like SAVI, SATVI, NDWI, DVI fail to explain the presence of soil nutrient based on spectral reflectance (Table 8).

Fig. 6
figure 6figure 6figure 6

a Indices of BGIR based on the reflectance of light A SATVI, B SAVI, C RDVI, D NPCRI, E ARVI, F BSI. b Indices of BGIR based on the reflectance of light G MAVI2, H MSAVI, I NDVI, J NDWI, K DVI, L NDMI. c Indices of BGIR based on the reflectance of light. M SCI, N EVI2

Full size image
Fig. 7
figure 7

Pearson correlation between the soil indices and the soil parameters [red line to blue line ranges from −1 to + 1 (i.e., −1, − 0.8, − 0.6, − 0.4, − 0.2, 0, 0.2, 0.4, 0.6, 0.8, 1 from red line to blue line)]

Full size image
Table 8 Descriptive statistics for all the 14 spatial indices
Full size table

Discussion

The study to analyze the physicochemical parameters using a trivial method like wet chemistry of soil and new method like remote sensing has been continuously used by agencies to determine the suitability of the soil (Xu et al. 2018a, b). To assess the suitability of the ex-situ conservation site at BGIR, this study has been carried out using the methods mentioned above (wet and remote sensing). The results obtained from the wet chemistry and spatial analyses have been discussed below.

Relationship between sulphur (S) and electrical conductivity (EC)

Results show that S and EC are strongly correlated (r = 0.97). According to Abdennour et al. (2019), the EC below 1ds/m is considered non-saline and does not have any impact on the loss of nutrients, soil microbe's activity consequently does not affect plantation or biomass. The available form of S is SO42− and in an ecosystem like botanical gardens, S cannot be a limiting agent. Also being an industrialized area, Noida is a polluted city, and polluted regions face acidic precipitation, which increases the SO42− availability in the soil and hence shows a positive correlation with the pH (Binkley et al. 1989). Sulphur can be made available from the oxides of iron and aluminium. At the same time, we have observed a weak interrelation between S and Fe (r = 0.23) and a moderate correlation of S with pH (r = 0.69). Still, the test is not significant at P < 0.05; the possible reason is the limited data on sample size. Similar kinds of results have been observed by the Likus-Cieślik et al. (2017)in their study.

Relationship between nitrogen (N) and organic carbon (OC)

Our study found that N and OC are robustly correlated because the carbon and nitrogen cycle are always linked to each other through production and decomposition. Carbon input will increase if nitrogen-fixing species increase the rate of litterfall and mycorrhizal production and also decrease the rate of detritus decomposition and humus formation leading to carbon accumulation in soil. Similar results have been found in the studies of Xu et al. (2018a, b).

Relationship between manganese (Mn) and organic carbon (OC)

Mn content is related to the texture of the soil; the coarser the soil, the lower will be the Mn content (Dhaliwal et al. 2022). A sufficient value of Mn is present in the studied soil. In previous studies, Mn has been found to be negatively correlated with the OC (Stendahl et al. 2017), while our results are a reversal of the previous studies, with a lower RSME (0.14) and high R2 (0.935) value which validates results in our study site. Mao et al. (2020) described that the soil OC is dependent on the soil texture; the coarser the soil, the lesser will be the organic carbon. OC content is low in the study site's soil because of the litter clearing from the floor, as this litter is being used to make compost, and this compost is further utilized for raising the samplings in the nursery. No direct cycling of carbon back into the ground's soil is observed.

Relationship between calcium (Ca) and pH

The Ca2+ ion concentration decreases as the pH is raised and hence shows significant negative robust interrelation with each other. The possible reason is phosphate ions at higher pH readily react with the Ca2+ and Mg2+ ions and make a less soluble form of them; as a result of this, Ca2+ content decreases in the soil (Johan et al. 2021), while past studies have shown a positive correlation between Ca2+ and pH (Koju et al. 2022).

Other correlations

Our study shows some significant Pearson correlation analysis results, which has not been considered for the MLRA analysis as N and Fe are not accepted as the predictor variable in our study; the only predictor variables we have used were EC, pH, and OC. In our research, Fe has shown a significant positive correlation with Cu (r = 0.94 at P < 0.05). A study by Benbi and Brar (1992) also supports our results. Results have also shown a positive correlation of Mn with N, having a Pearson correlation coefficient value r = 0.92 at P < 0.05.

Association of the soil parameters with the zones of the study area

The correspondence analysis between soil elements and the study site shows that Zone 3. C is highly associated with the EC and S. In contrast, Zone 5 is associated with the Zn and the taxonomy garden zone. It means Zone 3 can be planted with the trees, that need high sulphur conditions. In contrast, sulphur uptake studies have only been done in agricultural produce (Scherer 2001). Therefore, we concluded a research gap that needs to be studied carefully. This requires a study on those tree species with a high sulphur uptake requirement. EC aid in the availability of other salts to the plant, and Zone 5 can be used to plant trees with the high demand for Zn. Zn is essential for the growth of higher plants. Mainly fruit species are sensitive to Zn deficiency (Khan et al. 2022). Therefore, we can grow fruit species like Prunus persica, and Citrus reticulata, and under the greenhouse-controlled condition, we can grow Cydonia oblonga, Caryo illinoinensis etc.

Spatial index correlation with the soil nutrients

Different spatial indices correlate with the different soil nutrients. The NDVI value indicates leaf chlorophyll content and tell about biomass health conditions (Sahana et al. 2015). Here NDVI shows a positive correlation with both EC and S because EC is a predictor of soil properties like soil texture, organic matter content, Ca and Mg ions as well its cation exchange capacity and hence also a predictor of healthy biomass above the ground. So, the greater the NDVI value, the more the EC of the soil, while the S content of the soil increases with the temperature and organic matter content of the upper horizon of the soil (https://www.cropnutrition.com/). OM content is related to EC; therefore, dependency of EC on NDVI determines the S significant correlation with NDVI value (Bernardi et al. 2017; Sahana et al. 2020). NDVI is a most important indicator of soil carbon; while our study shows a weak correlation, which is not significant at P < 0.05, increasing the sample size can lead to a significant positive association of NDVI with carbon. EVI2 has shown a negative correlation with Mg in the soil at P < 0.05 level of significance. In previous studies, EVI/EVI2 has shown an association with soil moisture and land surface temperature (Jamali et al. 2011; Sedighifar et al. 2020), while there is no past study in which EVI2 has shown a correlation with the Mg content of the soil.

In our study, MAVI2 has shown a positive correlation with C and Mn. Leaf Area Index (LAI) and soil organic carbon (SOC) have been found to be strongly correlated (Kumar et al. 2017), and the LAI index is also correlated with the MAVI index (Zhu et al. 2014). Therefore, MAVI can be used as a good predictor of the OC content of the soil. Also, Mn shows a positive correlation with carbon content in the soil, as soil microbes use carbon for their oxidation reactions and provide electrons to Mn4+ to reduce it into Mn2+ and make it available to the plant (Morris and Gibson-roy 2019).

BSI is a tool for soil mapping and shows a strong positive correlation with soil pH and a negative correlation with Ca content. Bare soil is an indicator of high salt concentrations and the sandy nature of the soil (Sci et al. 2018). MSAVI index has also shown a negative correlation with Ca. The salts of Ca in the soil must be Ca (OH)2 and CaCO3 (Pal et al. 2000). BSI shows interrelation with the H+ concentration of the soil, while there is no supporting literature for the bare soil index and pH positive relation.

SCI, NDMI, MSAVI, and ARVI are all predictors of Zn concentration in the soil; SCI, ARVI, and NDMI index show a robust negative correlation with Zn (r =  − 0.89, r =  − 0.93, r =  − 0.88 respectively). The various form of Zn in the soil is Zn (OH)2, ZnO, and ZnCO3. The soluble forms of Zn decrease at pH 8 while {Zn (OH)2} is the dominant form at neutral pH; the species ZnSO4 and ZnHPO4 are the essential species for the availability of Zn in the soil solution (Alloway 2008). Our study site shows the pH around 8, which could be a possible reason for the less availability of Zn. Therefore, it shows a negative correlation with all the mentioned indices. On the contrary, MSAVI has shown a positive correlation with the Zn content in the soil (r = 0.92). Spectral reflectance is a tool to determine the soil P and Soil K status (Kawamura et al. 2011).

SCI is a predictor variable for the K in the soil as it shows a robust negative correlation with K concentration in the soil. The possible reason is the sandy nature of the soil, because the sand is negatively and significantly correlated with the K, as sand contains a very less amount of K bearing minerals, and SCI is also sensitive to soil background conditions, i.e., the sandy nature of the soil (Hewson et al. 2012; Saini and Grewal 2014).

NPCRI is used to quantify the chlorophyll content of the plant, though this index is also helpful in finding the B concentration in soil. In a study by Mouhtaridou et al. (2004), the B concentration in the media is inversely proportional to the chlorophyll content of the leaf and was found to be positively correlated with the NPCRI index. The B adsorption reactions in soil control the its concentration in the soil solution and hence control the B concentration available to plants. The factors that show a significant effect on B uptake by plants include soil texture, soil moisture, soil solution pH, and temperature. The B availability decreases with the increasing pH, and according to Wear and Patterson (1962), the soluble B is higher in a lower pH solution. Boron adsorption is noticed to be highest in the pH range of 3–9, and beyond this pH, the adsorption decreases (Steiner and Lana 2013). OC in the soil leads to more B availability (Gomes et al. 2017).

Fe is the most abundant element in the 'earth's crust; still, its availability is less because Fe gets oxidized in aerobic condition and form Fe oxides. In well-drained soil, the various abundant forms of Fe are Fe oxides, goethite, and hematite, while in poorly drained soil, the crystalline minerals (lepidocrocite, meghemite, and magnetite) or short-range ordered crystalline minerals (ferrihydrite and ferroxite) or non-crystalline precipitates (Cornell and Schwertmann 2003). Soil chromophores like iron oxides affect the spectral reflectance (Escribano et al. 2017), and in our study, the RDVI has shown strong interrelation with the Fe content in the soil. The redox potential and pH determine the availability of Fe to the plants; the acidic condition is favorable for the mobilization of the Fe (Colombo et al. 2013). At basic pH, the concentration of inorganic Fe is 10−10 M (Boukhalfa and Crumbliss 2002), which is assumed to be 104–105 fold lower than the necessitated for optimal plant growth (Römheld and Marschner 1986).

We recommend that the area should be introduced with (a) Nitrogen-fixing plants or bio inoculants, as nitrogen fixers, such as legumes or trees like Leucaena leucocephala, Dalbergia sissoo, Acacia nilotica, Pongamia pinnata, or any tree species belonging to the family Fabaceae, which grow in the agroclimatic zone of the upper Ganga plain region in India (Chaer et al. 2011). This will reduce the competition for N with other plant species. Increased diversity, which could lead to increased niche complementarity, is another strategy to improve nitrogen availability to the plants (Mulder et al. 2002), and C input will increase if nitrogen-fixing species increase litterfall and mycorrhizal production, decrease detritus decomposition and humus formation, and lead to C accumulation in soil (Binkley 2005); (b) Albizia lebbeck, L. leucocephala, Gliricidia sepium, and other mycorrhizal plants help to overcome phosphorus deficiency by solubilizing soil P through the release of organic acids and phosphatase enzyme (Billiash et al. 2019), and Trichoderma fungal sp. is also helpful in phosphorus solubilization (Saravanakumar et al. 2013; Bononi et al. 2020). (c) potassium-solubilizing bacteria, which can produce organic acids that aid in the solubilization of K-bearing minerals like Mica (Etesami et al. 2017). Mica, in the form of K can also be made available by applying composting technology, in which an acidic atmosphere converts unavailable K into a usable form (Nishanth and Biswas 2008). When used in tiny amounts, lime can help plants absorb more K by promoting root growth (Schroth et al. 2003; https://www.cropnutrition.com/). (d) Incorporation of Thiobacillium and Metallogenium bacteria into the field, since they aid in the dissolution of primary Fe minerals in the soil, as well as the availability and uptake of Fe by plants; as a result, we propose maintaining a slightly acidic pH (Colombo et al. 2013). (e) Use biosolids and acidifying fertilizers such as NH4(SO4)2, because Zn’ 's organic complex is soluble and mobile, increasing uptake by plant roots in the soil (Alloway 2008). To overcome B shortage, we recommend maintaining soil moisture and adding organic matter in the form of humic acid (Gomes et al. 2017), while the rest of the nutrients like Ca, Mg, Cu, Mn, S are found to be in the suitable range.

We would advocate transferring possible Indian forest kinds to this location by providing adequate climatic conditions inside conservatories to boost the scenic appeal and carbon sequestration of the area. The goal of introducing forest-type species into different BGIR zones is to promote the soil's OC. Because the SOC influences the availability of many other elements in the soil, we can enhance soil conditions by improving the SOC. The BGIR SOC (t/ha) is 2.61, but the SOC of the forest types is much greater; in this way, introducing forest types will improve SOC, and anytime the BGIR experiences leaf fall or a lack of flowering season for native plants, the introduced forest type species will bloom. To enhance aesthetic services, we need to keep each zone flowering and fruiting throughout the year. Botanists, researchers, foreigners, and other people will be drawn to the BGIR because of the year-round flowering and fruiting. Forest types are assigned to distinct zones depending on the few species that already exist in that zone. Forest kinds can be built without killing the species that already exist in zones in this way. The plan has been suggested to MoEF& CC and is now in process of implementation (Table 9).

Table 9 Proposed forest types in different zones of BGIR
Full size table

Conclusion

The present study focuses on using advanced technical methods to determine the suitability of the soil for successfully introducing the plants brought up from the different climatic zones for their conservation in botanical gardens, instead of just relying on the wet chemistry of the soil for checking the physicochemical parameters from time to time. The idea was to find the predictive variable and calculate the unknown value of the other dependent nutrient variable. The agricultural lands are in the vicinity of the garden on the Yamuna floodplain; the satellite-driven results provide the opportunity to understand the present status and the future demand of the soil nutrients for the cultivated crop. The study helps in understanding the specific nutrient enrichment of any zone or area; hence, the plant species can be cultivated accordingly. Based on our results, the farmers and policymakers can take the required actions for healthy crop production in the area without investing much into the process. The study clarifies the indices found to be the predictors of the soil nutrient data. Therefore, this study fills the research gap by predicting the crop nutrient quantity with the help of remote sensing by performing predictive regression between various indices and the physicochemical parameters. The results showed that a few indices like SAVI, SATVI, NDWI, and DVI failed to detect soil nutrient data. The research opportunity lies in developing new indices based on the spectral reflectance of the soil nutrients and vegetation at a different wavelength.