Application of MK trend and test of Sen's slope estimator to measure impact of climate change on the adoption of conservation agriculture in Ethiopia

Conservation agriculture is a crucial ecological issue impacting Earth and humans (Wilhite & Glantz 1985). It refers to water scarcity, severely impacting society's different segments, such as hydropower, agriculture, industrial, and water supply (Dracup et al. 1980). The crucial factors for assessing the conservation agriculture severity include period, location in absolute time, i.e., start and closure time points, areal coverage, and scale or force (Mishra et al. 2009). Climate and stream flow data from mountainous places have shown a pattern of increasing winter minimum temperatures over the last few decades and reduced snowfall compared to the total precipitation (Knowles et al. 2006). It has been indicated that decreased spring snow cover and snow depth affect summer stream flow (Rood et al. 2008). Studies in various countries have shown that snow significantly impacts the surface moisture accessible for runoff in the future (Maurer et al. 2003).

Climate change impacts are causing the biggest challenge in the world in terms of reduction in land productivity through reducing crop yield and increasing yield variability (Sehgal et al. 2020). This problem can be preserved by the adoption of conservation agriculture (Fleitas et al. 2020). Conservation agriculture is a field-level farming approach that fosters natural ecological processes to increase agricultural yields and sustainability by minimizing soil disturbance, maintaining permanent soil cover, and diversifying crop rotations (Juliana et al. 2020). It can also encompass natural resource management at the farm, village, and landscape scales to increase synergies between food production and the conservation and use of ecosystem services (Velu et al. 2020). As a context-sensitive management strategy, conservation agriculture can include diverse practices such as livestock and fodder management, improved fallows, agroforestry, watershed management, and community-protected areas (Su et al. 2021).

Several authors reported that conservation agriculture is an essential analysis to make long-term water management plans, particularly with regard to increasing agricultural yields and sustainability (Change 2014). In addition to this, biomass and residue data are regularly recorded and utilized for the planning and management of conservation agriculture (Hanssen et al. 2020). In this regard, a review of the literature revealed that several studies are conducted to understand the changing characteristics of flow using river flow data in developing countries, including Ethiopia (Benti et al. 2021).

Many approaches have been also proposed in the past to identify the trends in the adoption of conservation agriculture (Gabisa & Gheewala 2018). However, nonparametric approaches such as Mann–Kendall (MK) test and Sen's slope test have been preferred over parameter methods for biomass production in deserts due to the non-requirement of missing data, no-parameter estimated uncertainty, and simplified computation (Nguyen et al. 2022). The major drawback is that the significance of the test is severely affected by the presence of autocorrelation (Hu et al. 2020). Therefore, many researchers have proposed modifications to overcome the autocorrelation problem through conservation agriculture in order to increase yield (Malik et al. 2020). However, recent studies have revealed that while conservation agriculture may help us to reduce biomass phenomena such as the impact of climatic variability or land-use change (Alashan 2020), such localized variation may also result in a false trend that may impact the results of the MK test (Alhaji et al. 2018).

The other previous studies focused on the main socioeconomic factors that affect the adoption of conservation agriculture such as extension services, household head, high input cost, low input variability, possibly low biomass access, and greater size of land owned (Maina et al. 2020). There is a theoretical basis and limited empirical evidence to support the notion that climate variability affects the adoption and productivity levels of conservation agriculture. The impacts of climate change on the adoption of conservation agriculture have not been well studied (Abera et al. 2020). The variability of climate variables has a direct significant impact on the adoption of conservation agriculture (Hassen & Bantider 2020). However, there is not sufficient research to confirm this relationship as being broadly true. These relationships have not been specifically studied in adoption of the conservation agriculture practiced in Ethiopia.

The main objectives of this study were (i) to identify the long-term changes at different study areas assessing the adoption of conservation agriculture using Statistical DownScaling software; (ii) to compare the results of Statistical Downscaling software with the traditional MK and Sen's slope tests conducted in the study areas; (iii) and finally to identify whether adoption of conservation agriculture is increasing or decreasing in the study areas.

Furthermore, in order to give crucial information to smallholder farmers in rural regions, this study details how water erosion is one of the major issues with land degradation and constitutes a huge environmental hazard. The findings also helped us to explain how conservation agriculture might assist us to lessen the influence of events like climate variability or land-use change in the study areas on biomass. It also offers a comprehensive answer for Ethiopia's major issue with soil erosion brought on by excessive grazing, poor land management, and deforestation. Maintaining the cultivated land will help local residents by reducing the effects of high-intensity raindrops with high runoff potential. Additionally, it helps the smallholder farmer to pinpoint the key socioeconomic elements – extension services, family heads, high input costs, low input variability, potentially limited access to biomass, and ownership of larger tracts of land in the research areas – that influence the adoption of conservation agriculture.

The study was carried out in the Dire Dawa which is found in the eastern part of Ethiopia on the border of Awash plain on the highland just at the foot slope of the Dengego mountain between 9°27′ and 9°49′N Latitude and 41°38′ and 42°19′E Longitude (Solomon et al. 2020). It is astronomically located in the eastern part of Ethiopia 515 km away from the capital Addis Ababa (Gelata et al. 2022). It has a rugged and undulating mountainous topography ranging from 1,000 to 2,260 m asl and receives a total annual rainfall ranging between 410 and 850 mm and extreme temperatures of 34.6 °C and experiences frequent droughts, dry spells, and floods (Ayalew et al. 2017). The study area covers a total area of 3,143.8 km² and has a total population of 535,000 living in 38 rural and 9 urban kebeles (smallest administrative unit) (Anley et al. 2022). Agriculture (both crop and livestock production) is the main mainstay of the economy and subsistence mixed farming constitutes 93% of the total farm households in the study area (Mehari & Belay 2017).

The climate data (rainfall and air temperature from 1981 to 2021) for the present study were obtained from the NASA POWER (Prediction of Worldwide Energy Resources) project. The POWER project derives its data from the following NASA sources: World Climate Research Program (WCRP), Global Energy and Water Cycle Experiment (GEWEX), Surface Radiation Budget Project (NASA GEWEX SRB) and the Clouds and the Earth's Radiant Energy System and Assimilation Office at the Goddard Space Flight Center. Other researchers have used NASA POWER for their research and have justified its usefulness for metrological daily data to apply it in agro meteorology (Al-Adwan et al. 2022).

The temperature and rainfall data collected over the period were analyzed using statistical methods such as Mann–Kendall and Sen's slop. Several researchers have adopted this approach to understand the trends in the temperature phenomenon (Yaouba & Dieudonné 2022).

This test was originally developed by Sen for the performance of checking the statistical linear relationships (Agarwal et al. 2021). It is used to calculate the magnitude of trends in the long-term temporal data (Agarwal et al. 2021). Sen's slope is considered better to detect the linear relationship as it is not affected by outliers in the data (Ray et al. 2021). In this study, Sen's slope is applied to calculate the magnitude of the trend for temperature and rainfall data.

The positive (Q_i) indicates an increasing trend, while the negative Q_i values tell us that there is a negative trend in the temporal data. The unit of Sen's slope (Q_i) is the slope magnitude per year.

And then SDSM software was used to compare the observed and models climate scenarios of temperature and precipitation (Shakeri et al. 2021). For the calibration of the model, data from 1981–2005 was used, while 2006–2021 data was used for the validation (Sunil et al. 2021). The statistics for checking the performance of each of the models were the R2 and RMSE (Hassan & Hashim 2021). During the calibration, the downscaled parameters from the GCM which acted as the predictors were used iteratively to select the best predictor variables for the temperature (Shakeri et al. 2022).

According to the methodology for determining land degradation adopted by the UN for the calculation of the sustainable development goal's (SDG) indicator 15.3.1, land productivity on the basis of remote sensing data is one of the three sub-indicators (Shakeri et al. 2021). Then in order to estimate the biological productive capacity of the land, the sources of all the food, fiber, and fuel that sustain humans, we focused only on land productivity indicators (Sunil et al. 2021). Land productivity dynamics (LPD) data are provided by the Joint Research Council (JRC) as the default data for computing the final SDG 15.3.1 indicator (Wilby et al. 2002).

The MK and Sen's slope trend tests were individually applied at the different stations at seasonal, monthly, and annual time frame for the period 2021 at 90, 95, and 99% confidence levels (Ali et al. 2019). The methods were applied on monthly, mean annual, annual maximum, and annual minimum flow. Fitness of data was evaluated using the data of both stations at a 95% level of confidence (Kuriqi et al. 2020). The graphical results of SDSM were used to compare the observed and models (future) climate scenarios of temperature and precipitation.

According to the result, the models are significant and the observed and future scenarios was used for the assessment of trends. Parametric methods are more reliable compared to the nonparametric method (Gameda et al. 2022). However, parametric methods are applicable only when the data are normally distributed (Ayele et al. 2021). Distribution analysis of conservation agriculture data at both stations revealed that none of them is normally distributed, which means parametric methods cannot be used for assessment of trends. Therefore, nonparametric methods (MK and ITA) were used for the assessment of trends in the present study.

The result shows us the models are significant and the observed and future scenarios are similar through the year except those of April and August rainfall values. The model slightly over/underestimates some of the peak flows. This may have resulted from input data uncertainty and the weather or flow data quality used to input the model (Ayele et al. 2021). Some stations have missing weather data, which were left to be estimated and filled by the weather generator.

Indicators for the UN SDGs state that places are considered enhanced if their production is ‘improving,’ stable if it is ‘stable’ or ‘stressed,’ and ‘degraded’ if it is ‘moderately decreasing’ or ‘declining.’ According to the result (Table 1), earth analysis of the data from the years 2001 to 2021 revealed that 26.68% of the land area has increased productivity, 67.38% has remained stable, 5.93% has decreased productivity, and 0.02% has no productivity data. The Mann–Kendall and Sen's slope trend test of the 1981–2021 data reveals that the region experiences variability, a decrease in rainfall, and an increase in temperature. Yearly statistical error results indicate that the model performance is very good. However, it can be observed that the model overestimates and underestimates the stream flow at different times (Malik et al. 2021). The values were in high agreement compared to previous studies conducted in other regions of the world (Kumar et al. 2022). The values of the statistical indicators are in a close range to those obtained during calibration and validation.

The Sen's slope and MK trend test were individually applied to the study areas. The tests were applied at each month, annual mean, annual maximum, annual minimum, and seasonal flows for the period from the years 2001 to 2021 at 90, 95, and 99% confidence levels. The obtained results are presented in Table 2. Results show that there are positive (increasing) and negative (decreasing) trends in both stations (Ali et al. 2019). The highest negative trend can be seen in October while the highest positive trend can be seen in January at both stations.

The results of Mann–Kendall and Sen's slope test for the total maximum annual rainfall show a P-value of 0.307 which means there is no trend in the series or variability of rainfall and the negative Q_i values (−0.100) show that there is a decrease in the rainfall data. The slope of rainfall trend analysis shows that the rainfall graph for Dire Dawa station has a negative trend (Y = −0.001x + 60.22) as found by another researcher. For the total maximum annual temperature, the P-value is 0.01 which means there is a trend in the series or the temperature will continue increasing and the positive Q_i values (0.380) indicate an increasing trend.

Areas of land with increasing productivity from 2001 to 2021 are broken down by kind of land cover transition (Table 4). The productivity of the land increased from 40.00 to 84.86, 2.58 to 9.03, and 2.73 to 5.16 km² for tree-covered land, artificial land, and bare land, respectively. The productivity of the grassland and crops has decreased from 283.94 to 259.73 and 508.67 to 479.14 km², respectively. Additionally, there has been no improvement in the productivity of the land in the marsh and water bodies.

According to the result (Table 5), the tree-covered, artificial, and bare land cover show an improvement from 153.28 to 220.51, 3.87 to 9.18, and 16.16 to 18.51 with stable productivity, respectively.

By type of land cover transition, the area of land with stressed productivity (Table 6) with strained productivity, the amount of tree-covered and bare land increases from 1.29 to 1.44 and 0.00 to 0.15, respectively. With stressed productivity, the farmland, artificial land, water body, and wetlands exhibit little change. However, the productivity of the grassland has decreased from 2.73 to 2.43 due to stress.

According to the results (Table 7), tree-covered land, cropland, and artificial land all show increases with moderate declines in productivity from 0.15 to 0.23, 0.91 to 0.76, and 0.00 to 0.23, respectively. With a mild fall in production from 0.83 to 0.68 km², the grassland exhibits a reduction. Although productivity has moderately decreased, the wetland, bare land, and water body have not changed.

The above result (Table 8), the tree-covered and artificial land show an increase from 19.27 to 26.86 and 2.28 to 4.93 km² with declining productivity, respectively. The grassland and cropland show a reduction from 133.54 to 125.42 and 28.69 to 26.41 km² with declining productivity, respectively. But the wetland, bare land, and water body do not show any change with declining productivity. This is due to the variability and reduction of rainfall and increase of temperature in the region from the Mann–Kendall and Sen's slope trend test. Within a given ecosystem, primary productivity is affected by several factors, such as temperature, and the availability of light, nutrients, and water. Of those, increasing temperature and rainfall variability was the most significant factor in influencing the amount of plant tissue produced every year.

The findings show that conservation agriculture, a field-level farming strategy that promotes natural ecological processes to increase agricultural yields and sustainability by minimizing soil disturbance, maintaining permanent soil cover, and diversifying crop rotations, is becoming more and more popular. Additionally, future studies should take climate change projections into account in order to effectively estimate the intensity, frequency, and severity of droughts over a long time span (Ali et al. 2019).

Trend identification is one of the most important analyses in conservation agriculture adoption and climate change research, according to several earlier studies (Kuriqi et al. 2020). Our findings confirmed prior research showing that the adoption of conservation agriculture was more reliable for increasing agricultural yield in order to keep the sustainability policy for the 2023 goal (Wegari & Gelata 2021). The latter includes heavy tillage, which farmers already use to prepare the soil and clear crop wastes from agricultural fields for feed, fuel, and free grazing. Farmers till four to seven times annually.

This study demonstrated, among other things, that the goal for which water infrastructures were constructed is frequently at odds with extensive development and inefficient management of those infrastructures (Kumar et al. 2022). Effective management of water resources requires determining if low, medium, and high hydrological events have different trends or not in addition to establishing monotonic trends across a time range (Malik et al. 2021). Assumptions that may not always be accurate when considering socioeconomic aspects that influence the adoption of conservation agriculture include extension services, household head status, high input costs, low input variability, potentially limited access to biomass, and larger land ownership.

It is crucial to distinguish between the types of elements that have the greatest impact on flow alteration and availability in this study (Malik et al. 2021). On the other hand, the planning and implementation of the adaptation and mitigation measures that must be implemented in response to the alteration of water resources brought on by climate change and the occurrence of water scarcity events require additional time and money. The results of this study also showed that seasonality shift climate change can pose a risk to humid regions as well as desert climate zones (Ali et al. 2019). Consequently, more in-depth research in this area is required taking into account varied flow regimes and climates.

Risks associated with climate change's impact on the land cover could be decreased by enhanced conservation agriculture adaptation. Therefore, when applying conservation agriculture to agricultural, water, and food policies, a proper understanding of how climate change will affect its adoption is particularly helpful. For temperature and precipitation data, this analysis used NASA POWER, together with QGIS and Trends. Earth indicators and the research area's land production were examined. The trend result indicates that the region's land productivity has declined due to variability, less rainfall, and an increase in temperature. This suggests that conservation agriculture is being used less. However, with good planning and management of agricultural systems, these dangers might be greatly diminished. When there is a shortage of water, conservation agriculture should concentrate on efficient water use and water harvesting. In these conditions, maintaining the soil cover should be given less priority, and fertilizers should be utilized to prevent soil nutrient depletion. As biomass output rises over time, more organic matters may be added to the soil to improve its ability to retain water. This will increase soil cover. As a result, it is strongly advised that purposeful action be taken to create agroecosystems, institutions, and knowledge systems that can adapt quickly to changing conditions in order to avoid interruptions to the food supply.

The authors would like to acknowledge the research funds of the Emerging Agricultural Research and Reform Practice Project of MOE (2020-4-14) and the Project of local cooperation of Chinese Academy of Engineering (JS2020ZT12).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.