1 Introduction

Drought is described as a situation when there is insufficient precipitation or lack of enough water which affects vegetation greenness and health, plant mortality, reduced water levels, and other adverse ecological and/or biophysical conditions [1, 2]. Water-stressed plants typically close stomata to reduce water loss and prevent the development of excessively low water potentials, which significantly affect photosynthesis activity. This situation result in carbon stress, reduced growth, and greater susceptibility to insects, disease, and plant mortality [3, 4]. Rapidly the decrement in forest areas is negatively affected ecosystem function because there is a strong relationship between forest and ecosystem components. Changing environmental conditions and nutrients can affect the biota and related services provided by forests. Species disturbances, migration and extinction can happen as a result of forest degradation along with climate changes[5,6,7]. Therefore, gathering information required to monitor this natural resource is of utmost importance to maintain biodiversity conservation [8].

In recent decades, remotely sensed normalized difference vegetation index (NDVI) is used as a proxy for earth vegetation productivity that allows for continuous and long-term monitoring of information on the relationship between vegetation and climate and is one of the most widely used indicators in studies involving vegetation dynamics and ecosystem health across the world [9, 10]. Monitoring and assessing droughts in the semi-arid region at high resolution is also required the use of Remote Sensing based data and methods to fill a gap in in-situ data due to the drawbacks of using only ground-based data, such as inadequate spatial and temporal coverage of meteorological station data [11].

Changes in vegetation cover have attracted a focus of researchers to the assessment of vegetation response to changes in air temperature, precipitation, solar radiation, and sea surface temperature [12].

Precipitation variability is extra-ordinarily responsible for vegetation change in vegetation-atmosphere interactions when compared to the aforementioned climate variable due to the great role that precipitation plays in water availability. It is the leading factor controlling ecosystem structure and driving biological processes on over 40% of the earth’s vegetated surfaces [13, 14].

Moreover, examining the relationship between NDVI and climate variables helps us to recognize the most influential climatic variable in a particular region [15].

In Ethiopia, many studies have highlighted differences in NDVI changes and their response to climatic factors [16,17,18]. Therefore, those studies suggested that rainfall is the most crucial element to characterize the impact of drought on the environment in east Africa, particularly the semi-arid region of Ethiopia. In Ethiopia even though the relationship between remote sensing indices and precipitation was evaluated, the relationship between them has been not recognized at a small scale with improved resolution. In another way, many studies assessed agricultural drought using remote sensing indices from medium to large scale [15, 19, 20]. All previous studies reported that crop yield has declined as a result of drought severity specifically in the lowland region of Ethiopia. Although the importance of satellite indices is well recognized to assess drought impact on agriculture from medium scale to large scale, assessment of drought impact on forest resource at small scale has not been recognized.

Furthermore, several studies found that forests have been depleted as a result of anthropogenic activity in many regions of the country without considering drought impact on the forest ecosystem, using remote sensing Spatio-temporal forest change detection techniques [21,22,23,24]. Therefore, the novelty of this study is the time series analysis of drought severity assessment and its impact on the Yabello forest, using satellite drought indices, and assessing the relationship between forest and precipitation.

2 Materials and methods

2.1 Study area description

This study was conducted on the drought-prone region of Ethiopia, the Yabello forest, and its surrounding. The forest is located in the Yabello district of Borana Zone, southern Oromia. The total area of the study sites is 26,956.7 ha and its altitude range between 1438 and 2354 m. m.s.l [25]. The absolute location of this area lies between 4° 45′ 0" N to 4° 57′ 0" N and 38° 1′ 0" E to 38° 10′ 30" E (Fig. 1). The Yabello district has a bi-modal rainfall pattern with the main rainy season between March and May with high precipitation in April. The mean annual rainfall of the district is 587.2 mm. The mean annual temperature varies from 15 to 24 °C and shows little variation across the seasons [26]. At the Zonal level, the mean annual precipitation of the zone ranges from 352 mm in the southern part to 605 mm to the northern part of the zone. Geologically, most of the area which accounts for 40% is dominated by quaternary deposits, 38% basement complex formations, and 20% volcanic [27]. The region has a semi-arid savannah landscape, marked by gently sloping lowlands and flood plains vegetated predominantly with grass and bush-land. According to [27], this study area comprises three main soil types, 53% loam soil, 30% black clay, volcanic light-colored silty clay, and 17% silty. Five major land use/covers, namely; grassland, woody vegetation, cultivated, settlement, and bare land were identified in the district [22, 28]. Evergreen and semi-evergreen bush-land and thickets, rangeland dominated by shrubby Acacia, Commiphora, and allied genera, and dwarf shrub grassland to shrub grassland are also found in the area. The district has a total population of about 128,762 (64,692 Males and 64,070 Females) [29].

Fig. 1
figure 1

Map of Yabelo forest

Full size image

2.2 Data sources

In the present study, Landsat 8 OLI/TIRS with a 30-m spatial resolution (www.usgs.gov) was used as the data source to calculate NDVI and used to generate LULC, from the year 2014 to 2022. Since satellite images are application specific Landsat 8 is capable to detect forest cover changes at a small scale. These years were selected based on forest area a demarcated year and data availability. A total of twenty-two Landsat data was downloaded during the short rainy season (September to November) and for each month cloud-free satellites were downloaded in 2016 and 2020 more than 10% of cloud free data were omitted, and the mean of NDVI was used for each year (Table 1). This season was selected because of its higher rainfall variability [26]. Additionally, the most common statistical methods applied to time series of NDVI and precipitation, such as simple linear correlation or regression analyses; produce inaccurate results if seasonality is not taken into account [30]. Therefore, this research was undertaken to demonstrate a statistical modeling technique that would account for the seasonal timing effects of precipitation on the forest.

Table 1 Specifications of Landsat data used in the study
Full size table

The CHIRPS rainfall data was downloaded from the Famine Early Warning Network (FEWS). CHIRPS is a quasi-global (50ºS–50ºN) gridded product available from 1981 to near present at 0.05º spatial resolution (∼5.3 km) and at daily, pentadal, decadal, and monthly temporal resolution. It is developed by the U.S. Geological Survey (USGS) and the Climate Hazards Group (CHG) at the University of California [31]. Data were downloaded from 2014 to 2022 during the short rainy season (September to November) from the CHG web page (http://chg.geog.ucsb.edu/data/chirps/index.html). [32] used CHIRPS rainfall data for validating the drought monitoring tool in food-insecure regions of Ethiopia. In the same way, CHIRPS rainfall data was validated and showed the best performance than other rainfall data sources in Ethiopia [33, 34]. In this study, CHIRPS rainfall data was used for linear regression analysis between NDVI and precipitation.

2.3 Data analysis

2.3.1 NDVI calculation

During the early 1980s, the NDVI was defined and developed by scientists at NASA’s Goddard Space Flight Center, Greenbelt, Md. for monitoring vegetation health based on the difference between absorption and reflectance of green leaves of the red and near-infrared band of visible light, respectively[35]. The value of NDVI of each pixel is estimated by dividing the reflectance difference by the sum between NIR and the Red band. Its value range between − 1 and + 1, and lower values/near to zero represent stressed vegetation, negative values represent open water or high moisture content, and + 1 indicates healthy vegetation cover, respectively. Seasonal NDVI (September to November) was used in this study. To calculate NDVI indices from Landsat data, there are the pre-requested procedure which is guided by USGS, these include;

VNIR spectral radiance data were converted to top of atmosphere planetary reflectance using the reflectance rescaling coefficient provided in the Landsat Metadata file. The following equation is used to convert DN values to the top of atmosphere reflectance for the Landsat image.

$$\rho \lambda = {\text{M}}\rho\,{\mathbf{ \times }}\,{\text{Qcal}} + {\text{A}}\rho$$
(1)

where ρλ = TOA planetary reflectance without correction for the solar angle.

Mρ = represents a band-specific multiplicative rescaling factor from the Metadata Reflectance Mult-Band X where X is the band number.

Qcal is quantized and calibrated standard product pixel value (DN).

Aρ represents the band-specific additive rescaling factor from Metadata Reflectance-Add-Band-X, where X is the band number. The next procedure is correcting the reflectance value with the sun angle.

The formula for correcting this value is;

$$\rho \lambda = \frac{{ {{\rho \lambda }}}}{{\cos\theta {\textit{SZ}}}} = \frac{{ {\rho \lambda }}}{{\sin\theta {\textit{SE}}}}$$
(2)

where ρλ = TOA-planetary reflectance.ɵSE = local sun elevation angle, the scene center sun elevation angle in degrees is provided in metadata (sun elevation).

The local zenith angle is calculated as Eq. (3)

$${\mathbf{90}}^{{\underline { \circ } }} - {{\uptheta }}{\textit{SE}}$$
(3)

To obtain more accurate results, its unit should be converted from radian to degree. These procedures were carried out for all bands used in this study which operate in the optical region of the electromagnetic spectrum (bands 4 and 5). Finally, after these corrections were completed bands were extracted into the study area by using an ArcGIS extraction tool. Then NDVI was generated by using an ArcGIS raster calculator.

$${\hbox{NDVI}} = \frac{{{\textit{NIR}} - {\textit{RED}}}}{{{\textit{NIR}} + {\textit{RED}}}}$$
(4)

2.3.2 Vegetation condition index

VCI is useful for vegetation condition assessment, as it assesses changes in NDVI through time since vegetation is water-stressed due to water deficiency such as during drought as stated by [36].

According to Kogan-VCI is measured as a percentage with values ranging between 0 (Lowest) and 100 (Highest) as in Eq. 5 with values equal to or below 40% considered as drought to varying degrees of severity (Table 2).

Table 2 VCI and Drought severity class
Full size table

Therefore, in the present study drought severity during the growing season (short rainy season) was measured using the NDVI-based Vegetation Condition Index (VCI) for 8 years, for the same period using the ArcGIS raster calculator tool as in Eq. (5).

$$VCI = 100 \times \frac{NDVI - NDVImin}{{NDVImax - NDVImin}}$$
(5)

where

NDVI = the value of NDVI at the time of observation.

NDVI min = Absolute minimum NDVI value of 8 years.

NDVI max = Absolute maximum NDVI value of 8 years.

2.3.3 Forest cover change

To access drought’s impact on the forest, forest cover change was evaluated during the study period. Land use/land cover of 2021 was used as the reference to evaluate the variability of forest cover change as the result of drought. This year was selected because a large portion of the region was covered by forest and to minimize the contribution of settlement and agricultural activity in the result of this finding.

Recently, settlement and agricultural activity are rapidly increasing. Therefore, in this study settlement and others human activities were maxed out from the study area to identify drought impact on the forest region by omitting the impact of anthropogenic activities. Land use/land covers were classified based on the NDVI classification range as used by [37] at twelve districts of Ethiopia, which belong to the same drought-prone area. These NDVI ranges were >  =  − 0.3 <  = 0.084 is a water body, > 0.084 < 0.18 is bare land, >  = 0.18 < 0.3 is sparse vegetation, and >  = 0.3 < 0.6 is dense vegetation. Based on these NDVI ranges spatial–temporal forest covers were evaluated over 8 time series years using ArcGIS image analysis.

2.3.4 Correlation analysis between precipitation and NDVI

A simple linear regression model was applied to examine how the precipitation–NDVI, relate to each other based on their mean values using “Microsoft excel”. The mean values for both variables for each year were evaluated.

Besides, the Pearson correlation matrix was applied to evaluate the relationships of an index. The linear regression model was computed as follows.

$${\text{Yi}} = \beta 0 + \beta {\text{1Xi}} + \varepsilon {\text{i}}$$
(6)

where

Yi = NDVI, for the ith period, Xi = precipitation, β0 + β1Xi = linear relationships between Yi and Xi, β0 = mean of Yi when X = 0 (intercept), β1 = change in the mean of Y when X increases by 1 (slope), εi = random error term.

The results of simple linear regression or Pearson correlation can be positive or negative. This ranges between 0 and + 1. Regression or Pearson correlation values close to zero indicate no relationship between the indices. However, if one index increases as the other index (e.g., NDVI) decreases, then the relationship is negative (Fig. 2).

Fig. 2
figure 2

Methodological flowchart of the study

Full size image

3 Results

3.1 Drought severity

The vegetation condition index revealed drought years over the study area during the study period.

Accordingly, the most drought years were detected in 2015, 2017, 2019, and 2022 years as it’s shown in Fig. 3. Spatially, in 2015, 2017, 2019, and 2022 years, severe drought covered 177.5 ha, 537.9 ha, 69.7 ha, and 80.6 ha corresponding to the aforementioned years (Table 3). Northwest, central, and southeast of the forest were frequently affected as the result of drought (Fig. 3). Also moderate drought was detected in these years over the spatial magnitude of 2,994.49 ha, 7,116.2 ha, 7,359.7 ha, and 4,449.6 ha corresponding to above years. In these years, only 1,397 ha, 4420.7 ha, 8,563.7 ha, and 14,381.4 ha were not affected by drought (Table 3).

Fig. 3
figure 3

Spatial map of drought severity class

Full size image
Table 3 Spatial extend of drought class during 2014–2022
Full size table

The wettest years were also identified during the study period. For example, in 2014 even though a severe drought occurred over 80.1 ha, a large portion of the region (15,783.7 ha) remained normal (Table 3). In 2016, severe drought covered only 7.6 ha, and moderate drought covered 665.3 ha whereas, a larger area (19,048.7 ha) remained without being affected by drought. In 2018, severe drought covered 43.6 ha, moderate drought covered 1,591.2 ha of the Yabalo forest, and 15,581.2 ha was not affected by drought (Table 3). Similarly, in 2020 and 2021, severe drought did not occur in the Yabalo forest. In 2020, only 37.9 ha belonged to moderate drought stress, and the maximum spatial extent of the wettest forest condition (24,644.2 ha) was detected this year. In 2020, 37.9 ha moderate drought might be indicated that forests are recovering from the drought that occurred in 2019 (Table 3). Even though severe drought was not recorded in 2021, there was an implication to drought occurrence from analyzed data. For instance, the maximum normal forest coverage in 2020 (24,644.2 ha) was decreased to 18,403.5 ha in 2021. Also, the mild drought of 269.9 ha in 2020 increased to 6,568.5 ha in 2021. In2021, moderate, mild, and normal drought categories covered 4.3 ha, 6,568.5 ha, and 18,403.5 ha of land respectively (Table 3). Following 2021, drought severity was detected in the region in 2022. In 2022 severe, moderate, mild, and normal forest land mass expanded over 80.6 ha, 4,449.6 ha, 6,050.3 ha, and 14,381.4 ha respectively. In all of the detected drought year, the northwest, central, and southeast of the forest was the most affected portions of the Yabelo forest (Fig. 3).

3.2 Impact of drought on forest

NDVI-based LU/LC revealed that forest cover has been not stable during the study period. For instance, as indicated in Table 4 in 2014 dense forest, sparse forest, and bare land covered 305 ha, 19,660 ha, and 4,995 ha respectively. But following 2014, in 2015 the dense forest was decreased while sparse forest and bare land were increased. This is maybe because of the drought that occurred in the Yabello forest in 2015. The dense forest was decreased by 88 ha and bare land was increased by 39 ha as compared to 2014. Thus, 88 ha of dense forest was converted to sparse forest and 39 ha of sparse forest was converted to bare land as the result of drought. The result found that in 2016 the maximum dense forest was detected during the entire study period. Accordingly, 424 ha of dense forest was detected and increased by 207 ha. Sparse forest and bare land decreased this year. This indicated that sparse forests were rehabilitated and changed to the more dense forest while bare land was rehabilitated into the sparse forest during this time. Contrary to 2016, in 2017 dense forest was significantly decreased by 231 ha. Though the sparse forest was increased, the maximum bare land area was registered this year. As indicated by the above VCI result maximum extreme drought was registered in 2017. As the result, 100 ha of sparse forest was lost to drought this year as it’s indicated in Table 4. In 2018 both dense forest and sparse forest increased and bare land decreased. This indicated that because the study area experienced normal climatic variables; vegetation was recovered during this time. Hence, the dense forest was increased by 179 ha, the sparse forest was increased by 268 ha and the bare land was decreased by 457 ha. This study found that drought that occurred at every interval of 2 years (2015, 2017, and 2019) brought significant change to forest resources. For instance, in 2019 dense forest and sparse forest decreased by 140 ha and 163 ha respectively as compared to 2018, and bare land increased by 303 ha. Following 2019 the two consecutive years thus 2020 and 2021 were experienced the normal year. Hence, both dense forests and sparse forests increased while bare land decreased. Contrary, in 2022 both dense forest and sparse forest decreased whereas bare land increased.

Table 4 Forest covers change
Full size table

The spatial mapping of the impacts of the major registered drought event in the study area during the study period showed that vegetation activity across the entire study area decreased, as drought occurred over the region (Fig. 4). Thus, the forest has not stables during the study period as the result of climate change. In all of the detected drought years, the dense and sparse forest was diminished through the Northwest, central, and southeast of the forest (Fig. 5).

Fig. 4
figure 4

Dense forest, Precipitation, and Severe drought trend during the study period

Full size image

To explicitly, illustrate the impact of drought on forest cover, the below histogram was used to indicate the trend of three variables that were purposively selected (Fig. 4).

3.3 Relationship between precipitation and forest

In this study, the relationship between precipitation and NDVI were examined to recognize how precipitation controls the function of environmental resource, particularly forest. The result of this finding showed that there is a positive correlation between precipitation and NDVI. The result of R2 as shown in Fig. 6 is 0.49. Thus, 49% of NDVI that indicates vegetation condition can be determined by precipitation since the value of R2 was 0.49 in this study (Fig. 6). The regression analysis result also showed that the relationship between precipitation and NDVI is positive, (R/P = 0.69/0.05) and statistically significant. Thus, the results indicated that when the precipitation increases, NDVI also tends to increase. When precipitation deficit occurred in this Yabello forest region vegetation cover and greenness became decreased and vice versa. For instance, in 2018 mean precipitation of the short rainy season was 195 mm and highest as compared to other years during the study period. Also, vegetation greenness was high and the maximum NDVI value (0.27) was registered this year (Fig. 6).

4 Discussion

Northern, eastern, and southern regions of Ethiopia have been frequently affected by drought. The frequency of occurrence of drought is becoming beyond the loss of livestock and crop failure and resulting in forest degradation. As indicated in the VCI result, drought has occurred in the Yabello forest area nearly every two years. This finding is in line with several studies that projected drought occurrence in the lowland region of Ethiopia to have occurred every two to five years [38, 39]. Also, the Study conducted in the “Yabello and El-Woye area reported that Yabello districts had an extremely severe drought with SPI value—2.28 at a two-month timescale which belongs to the season of this study (September to October) in the 2015 drought year from 1987 to 2017 [40]. Similarly, the drought that occurred in 2017 resulted in severe damage to assets and exposed hundreds of thousands to famine and costs for relief as highlighted by [41]. According to the report of [42], the eastern Horn of Africa region has experienced drought conditions following consecutive poor seasons since late 2020 leading to significant impacts on production, vegetation, water resources, food insecurity, and livelihoods. In line with this report, the VCI result of this study indicated that vegetation greenness was beginning to decrease as precipitation deficit was reported for three consecutive seasons. Hence, drought severity appeared in the Yabello forest. This report stated the drought years of 2017 and 2019 in the southern region of Ethiopia which confirmed the result of this study.

Because of climate extremes like drought, the forest can be converted to bare land or the more dense forest can be sparsely distributed. Several studies reported as forest cover declined as a result of drought. For instance, a study conducted on forest cover change and climate variability in southern Ethiopia determined that the forest cover was decreasing from 10,020.3 ha to 8,851.2 ha, between 1986 and 2018 respectively as a result of climate change [43]. A similar study also found that forest cover was changed as a result of rising temperatures from 1990 to 2021 in Turkey [44]. The negative impact of rising of temperature along with drought is not only confined to the forest. but, it can cause migration/diminishing of species [5, 6] Likewise, in 2015, 2017, 2019, and 2022 drought brought the loss of many life stock productions in the Borana lowland region of Ethiopia. The drought resulted in the loss of 231 ha of dense forest and the loss of 100 ha of sparse forest in 2017 as compared to 2016 in the Yabello forest. The drought resulted in plant mortality as suggested by [30, 45]. But, based on vegetation types and structure the ways forests respond to drought are different. For instance, in this study dense forest was converted to sparse forest, and sparse forest was lost to drought. This study also found that drought seriously diminished dense forests in the 2017 year (Table 4 and Fig. 5). In line with this finding, a study conducted on Taltele range land which is a neighbor of this study found that the larger change in land observed in forest, grassland, and wetland was detected during the period from 1995–2000 and 2015–2019, this is due to climate change impact happened in Taltele rangeland [21]. A similar study found that forest cover has shown variability from 1986 to 2018 in the Kafa zone using NDVI indices [43]. Also, [46] observed that vegetation activities decreased as drought appeared over Southeastern Europe using SPEI.

Fig. 5
figure 5

Map of forest cover change

Full size image

To conserve natural resources, knowing the relationship between drought indices plays a great role [47]. In the semi-arid region, precipitation is the most important one of the climate elements to recognize the biophysical environmental function. As a result, several studies examine the relationship between precipitation and NDVI to recognize how vegetation is stressed as a result of precipitation deficit [4, 11, 48]. Commonly, they found that positive relationship between precipitation and NDVI. Similarly, this study confirmed that precipitation variability is extra-ordinarily responsible for vegetation change in the Yabello forest, due to the dominant role that precipitation plays in water availability which has been considered as the principal factor controlling ecosystem structure and driving biological processes in vegetation. Thus, when precipitation is a deficit for the plant, the plant closes its stomata to conserve water. Hence, photosynthesis is no longer functions and plant mortality can be occurring (Fig. 6).

Fig. 6
figure 6

Relationships between NDVI and Precipitation

Full size image

5 Conclusion

Climate extreme like drought is challenging the globe in multi-direction. Its negative impact adversely affects ecosystem function. Particularly its impact on forest resources is attracting the attention of many scholars, and decision-makers also required information regarding drought characteristics. This study proposed spatial and temporal analysis of drought evolution in the Yabello forest area in the lowland region of Ethiopia from 2014 to 2022. The results showed the usefulness of incorporating forest cover change in assessing the Spatio-temporal variability of drought severity in the dry-lands. Remote Sensing-based vegetation time series indices were used, as complementary to in-situ data indices. NDVI-based forest cover change and VCI Indicators were used for monitoring changes in vegetation conditions and drought severity during the growing seasons (September to November) from 2014–2022.

Results showed high temporal and spatial variability in vegetation cover between drought and non-drought conditions in this study. The associated negative impact of droughts on the vegetation resulting in limited forest cover was further confirmed by results from this study. Severe drought occurred in 2015, 2017, 2019, and 2022 and forest cover was highly affected during these years. The wettest years were 2014, 2016, 2018, 2020, and 2021.

Simple linear regression analysis that considered precipitation as the independent variable and forest greenness as the dependent variable confirmed that half of the forest degraded was because of precipitation deficit. Due to a lack of precipitation, forest has been degraded every two years in this Yabello forest. This study highlighted drought impact as the most climate phenomenon that causes forest degradation in the semi-arid region of Ethiopia which requires policy actions toward climate change adaptation options. Moreover, future research should be integrating remote sensing metrics with other factors like socio-economic and biophysical data to assess drought magnitude and intensity in the study area.