Climate Change Adaptation in Agriculture: The Case of Small Southern Mediterranean Country

In this study, we examined the effect of climate change on the incomes of farmers in a southern Mediterranean country. We proposed that crop insurance could be potentially used as a means to adapt to climate change. Using panel data for Tunisian regions, we were able to highlight the 10 important effects of climate change on crops yields by considering two scenarios of the 11 Representative Concentration Pathways, namely RCP 4.5 and RCP 8.5. In the long term (i.e., in 12 2050 and 2100), we expect increasingly frequent heat waves to occur, leading to a rise in droughts 13 for all regions of Tunisia. We therefore recommend that farmers seek to insure themselves against 14 the risks of drought and flood to their crops, because we feel this may be an attractive device for 15 compensating them for any potential losses of income.


35
Climate change now appears to be a proven phenomenon, and it has been found to be 36 linked, at least in part, to emissions from human activity contributing to the greenhouse effect. are being seen at the regional level. Climate change prompts us to question how we manage 42 the natural resources of our planet, and it puts our production methods to the test, especially 43 for the agricultural sector.  These risks constitute a problem for the development of agricultural assets, which are needed 56 to increase the incomes of farmers and reduce rural poverty (Dilley et al., 2005). 57 Over time, farmers have mastered cultivation techniques that are well suited to the local 58 climatic conditions, yet climate change implies that these will need to be adapted, which in 59 turn requires the rapid development of innovative economic and social practices. The use of 60 insurance may be one such practice. Mitigation and adaptation measures in agriculture could 61 rely on effective levers like specific insurance contracts to limit the economic impact of 62 climate change. Insurance companies are now on the "frontlines" of climate change, which is 63 turning out to be more of a threat than an opportunity. They can therefore help manage 64 climate risk through products and services for agricultural activity to help limit the financial 65 cost of harmful climatic events (Kath et al., 2018). 66 The research tends to suggest that climate change will affect the future incomes of farmers 67 in the form of unreliable crop yields. One way to adapt to this situation could comprise taking 68 out an insurance policy that has been developed for the agricultural sector and the regional 69 heterogeneity of climatic risks. Moreover, this work highlights the important implications of 70 future episodes of drought for Tunisia by 2050-2100. The regional climate models (RCMs) of 71 the EURO-CORDEX project constitute the basis for our analysis, which is based on the 72 average emissions scenario RCP 4.5 and the pessimistic emissions scenario RCP 8.5.

73
Mediterranean agriculture is highly vulnerable to climatic phenomena such as drought, 74 excessive or insufficient rainfall, heat waves, and so on, and such climatic phenomena    In recent times, we have seen increasingly frequent extreme weather events, and these have  Having insurance companies help manage the risk of extreme climate-related events will 121 require encouraging insurers to offer adequate cover against these risks and farmers to take 122 them up. The insurance market is relatively new in Tunisia, which is a small country on the 123 southern Mediterranean coast that is characterized by cold, rainy winters and hot, dry 124 summers. Insurance companies that specialize in agriculture mainly insure against crop losses 125 caused by natural disasters like drought, hailstorms, insects, frost, and so on. The setting of 126 premiums based on risk management should be at the core of insurers' concerns given the 127 increasingly frequent and severe natural disasters that have been seen in recent years. For the 128 insurance industry, income from premiums often lags behind the growth in claims. What is 129 more, with unpredictable but economically harmful events, merely considering historical data 130 does not, in itself, lead to suitable insurance premiums (Tucker, 1997).

131
In the agro-economic literature, climate change has been shown to have a significant 132 potential effect on agricultural production, and the relationship between the two phenomena 133 has been the subject of several studies (Wheeler & Von Braun, 2013; Ruminta, 2016; 134 Suryanto et al., 2020). Farmers will tend to see their incomes decline because of climate 135 change. Indeed, when faced with infrequent but damaging natural events, farmers will suffer 136 significant financial losses. Research has posited that the best way to adapt to these 137 unpredictable, extreme climatic risks is to seek insurance through government programs.   153 We sought to study the effects of global warming on crop yields and their variability.

154
Econometric modeling therefore investigated the relationship that may exist between yield 155 variability and meteorological factors. 156 Villavicencio et al. (2013) found that the temporal behavior of crop yields is not constant 157 due to its strong dependence on climatic factors, so any estimates based on a level series will 158 be spurious. Panel stationarity tests were carried out before proceeding with the study to 159 investigate the dynamics between the different variables in the study. This can be estimated 160 using the cointegration technique, which has two underlying advantages: First, it overcomes 161 the difficulty associated with the narrowness of a time series, and second, it is more powerful 162 than traditional time series tests (Banerjee, 1999).

164
The use of unit root and cointegration tests for econometric panel data offers a genuine 165 advantage for dynamic models. Much research work has shown that there is a considerable 166 improvement in the power of unit root tests when using panel data.

167
In recent decades, several researchers-such as Levin, Lin, and Chu (2002) Where we assume the absence of a temporal effect and inter-individual dependencies, that 173 is, The second category of tests allows the heterogeneity of autoregressive roots under the 176 alternative hypothesis to be established. These tests seem better suited to our regional panel  cointegration relationships. This test is based on the following long-term relationship: and adjustment rates under the alternative hypothesis, which appears to be particularly 208 relevant to our model given the significant regional differences observed in yield losses.  The EURO-CORDEX project uses temperature and precipitation as the basis for 242 simulating 14 regional climate models. The Tunisian National Meteorological Institute, as an 243 associate of the EURO-CORDEX project, carries out climate forecasts for the entire territory.

244
The changes in temperature and cumulative rainfall were calculated for two emission 245 scenarios, namely the average scenario RCP 4.5 and the pessimistic scenario RCP 8.5.

284
To characterize climate change, we used temperature (minimum and maximum) and 285 precipitation as climate attributes. We used the cumulative precipitation for September to 286 March, which is the rainy season in the southern Mediterranean.   between Corp_Yield_los, Av_temp, Min_temp, Max_temp, and Cum_pluv.

309
As indicated in Table 2, the Pedroni and Kao tests indicate that the null hypothesis of no 310 cointegration can be rejected for all regressors at a 1% significance level (five against three).

311
This implies that there is a long-run relationship between yield losses and climatic factors.

313
Given the existence of a long-term relationship among the variables, we estimated equation 314 4 using the FMOLS and DOLS methods. Table 3      The agricultural sector is relatively vulnerable to climate change, so it will be heavily 346 affected by it. One of the most damaging aspects of climate change for Tunisia is increased 347 drought, which will harm food production and consequently affect the incomes of farmers. theory predicts the premium will increase as the risks from climate change increase.

362
The projections made based on the cointegration relationship for yield losses show that 363 there will be a significant increase in crop losses. Climatic disasters are infrequent, however, 364 and this limits the ability of insurance companies to assess and determine the level of risk.

365
Higher prices are therefore a likely consequence of the increased uncertainty facing insurance 366 companies.

367
Without government intervention, these potential long-term losses will result in higher 368 insurance premiums and possibly the withdrawal of insurance companies from this market.

369
The insurance system will therefore vary in terms of viability and efficiency, but the 370 introduction of public-sector insurance makes it possible to overcome the shortcomings of the 371 free market in providing cover for extreme climatic risks. In some countries, agricultural 372 insurance is subsidized by governments through national production protection programs.

384
The cointegration model estimation for the panel data allowed us to observe regional 385 disparities in the effect of climate change on the productivity of agricultural land. These 386 results, when related to climate change scenarios, should enable farmers to better adapt and 387 better plan for the possible risks, such as by taking out crop insurance.

388
The projections based on the two scenarios RCP 4.5 and RCP 8.5 show an increased risk of 389 yield losses, especially in some higher risk regions. This suggests that the demand for 390 agricultural insurance in Tunisia could increase in a future that seems likely to be 391 characterized by greater and more frequent climatic events. Mitigation and adaptation 392 measures for climate change should therefore be taken to cope with the potential yield losses.

393
In Tunisia, heat waves in recent years have given rise to fires that have affected crops, 394 resulting in damage to farmers' incomes, the environment, and biodiversity. In future, such 395 phenomena may intensify and cause some farmers to give up on agricultural production.