Environmental Erosion Research

fa برآورد رطوبت خاک بر اساس تولیداتNDVI و LST مودیس در مناطق فاقد داده Soil moisture estimation based on MODIS NDVI and LST productions in areas with no data مدلسازی و تحلیل زمانی و مکانی رخداد انواع مختلف فرسایش محیطی مستخرج از پایان‌نامه / رساله / طرح پژوهشی برآورد رطوبت خاک سطحی برای مدیریت بهینه منابع آب و خاک ضروری است. رطوبت خاک سطحی، متغیری مهم در چرخه آبی طبیعت است که نقش مهمی در تعادل جهانی آب و انرژی و فرآیندههای هیدرولوژیک، اکولوژیک و هواشناسی دارد. رطوبت خاک بهدلیل تغییرپذیری ویژگیهای خاک، توپوگرافی، پوشش گیاهی و پویایی نیوار در زمان و مکان تغییر میکند. اندازهگیری رطوبت خاک، بهطور مستقیم با استفاده از اندازهگیریهای میدانی مانند نوترونمتر و <a href="#_ftn1" name="_ftnref1" title="">[1]</a>TDR یا بهطور غیر مستقیم بهوسیله توابع انتقالی و یا سنجش از دور انجام میشود. از آنجا که اندازهگیریهای میدانی معمولا در پهنههای وسیع هم هزینهبر و هم زمانبر و گاهی نشدنی میباشد، برای برآورد رطوبت خاک در مقیاسهای مکانی بسیار بزرگ، میتوان روشهایی همچون سنجش از دور را بهکار گرفت. این تحقیق با هدف بررسی امکان برآورد رطوبت لایه سطحی خاک با استفاده از تصاویر سنجنده مودیس و مقایسه آن با دادههای زمینی انجام شد. در این مطالعه، رطوبت خاک در عمقهای صفر تا سی سانتیمتر با استفاده از رابطه بین شاخص گیاهی ماهوارهای (<a href="#_ftn2" name="_ftnref2" title="">[2]</a>NDVI)، دمای سطح زمین (LST<a href="#_ftn3" name="_ftnref3" title="">[3]</a>) و رطوبت مشاهدهای خاک در مقیاس منطقهای با تفکیک مکانی یک کیلومترمربع برای سالهای 2021 و 2022 برآورد شد. ضریب تعیین R2 و AIC معادلات رگرسیون بهترتیب 78/0 و 2/166 بهدست آمد که نشان می‌دهد رویکرد برآورد بر اساس داده‌های NDVI و LST مودیس، مناسب بوده و می‌تواند برای تخمین رطوبت خاک طی سال‌های 2007 تا 2022 استفاده شود. <div>  <hr align="left" size="1" width="33%" > <div id="ftn1"><a href="#_ftnref1" name="_ftn1" title="">[1]</a> Time Domain Reflectometry</div> <div id="ftn2"><a href="#_ftnref2" name="_ftn2" title="">[2]</a> Normalized Difference Vegetation Index</div> <div id="ftn3"><a href="#_ftnref3" name="_ftn3" title="">[3]</a> Land Surface Temperature</div> </div> <div>  <hr align="left" size="1" width="33%" > <div id="ftn1"></div></div> 1- Introduction Soil moisture can be considered in the control of desertification, agricultural activities, watershed management and optimal management of water resources.   Since the country of Iran is facing many problems in these fields, the expansion of studies in the field of accurate estimation of soil moisture becomes important (Mehrabi et al., 2019). The GLDAS system may have a high error compared to the measured data in some areas. Therefore, it is necessary before the data and results of this product are used as a decision-making tool in the region.  The quality of these data should be evaluated locally using ground-measured data (Polo et al, 2016 & Sanchez-Lorenzo et al, 2013 & Zhang, 2019).In this study, Terra MODIS data was used to estimate soil moisture due to higher spatial resolution (1 km). Due to the fact that it is difficult to estimate the humidity time series in the field, and radar remote sensing methods produce humidity maps with low spatial resolution. Therefore, in this study, a new method was introduced to prepare a soil moisture map with higher spatial resolution based on NDVI and LST MODIS products.The purpose of this study is to estimate soil moisture in Jiroft city using the products of the Morris sensor and NDVI and LST indices. Considering that Jiroft plain is one of the agricultural poles of Iran. Estimating the time series of soil moisture and then providing a drought index based on soil moisture is a useful method for investigating agricultural drought in the study area. 3- Results LST and NDVI have high relative importance in arid regions (Park et al, 2016). In the study area, an algorithm based on remote sensing was used for soil moisture time series due to the lack of access to soil moisture time series. Moody's LST and NDVI products with a resolution of 1 square kilometer were used during 2007-2022. Then multiple linear regression was created using OLS method between soil moisture observations (2021 and 2022) and NDVI and LST time series data. Based on the results, both independent variables NDVI and LST were significant at the 99% level) p-value<0.01) .VIF values were less than 10. Therefore, there is no linearity between the independent variables. The standard error is small, which indicates that the estimated value is exactly the true value. The value of t statistic of two variables is higher than 2, which is considered statistically significant. In general, R2 = 0.78 and it shows the accuracy of soil moisture estimation method. R2 was obtained as 0.74 and 0.8 in the years 2021 and 2022, respectively, which indicates the accuracy of the predicted values. 4- Discussion & Conclusions Estimation of spatial and temporal changes of soil moisture is an important issue in low data areas such as Iran.  We proposed a multiple regression model based on Modis NDVI and LST to obtain surface soil moisture at a regional scale. The results showed that the multivariate linear regression method can be used to estimate soil moisture products with high resolution in areas with little data in the surface layers of the soil. Park et al. (2016) stated that LST and NDVI have high relative importance in arid regions. In the studied area, according to previous studies (Bai et al, 2020 & Wang et al, 2007), an algorithm based on remote sensing was used for soil moisture time series due to the lack of access to soil moisture time series became. Moody's LST and NDVI products with a resolution of 1 square kilometer were used during 2007-2022. Then multiple linear regression was created between soil moisture observations (2021 and 2022) and NDVI and LST time series data using OLS method.  Based on the results of simulated soil moisture changes and the OLS method in estimating soil moisture (0-30 cm), both independent variables NDVI and LST were significant at the 99% level (p-value <0.01)). VIF values are less than 10 Therefore, there is no linearity between the independent variables. The standard error is small, which indicates that the estimated value is exactly the true value. The value of t statistic of two variables is higher than 2, which is considered statistically significant. Overall, R2 was 0.7  And it showed the accuracy of the soil moisture estimation method, which is consistent with the results of (Khanmohammadi et al, 2008 & Lin et al, 2015). رطوبت خاک؛ سنجش از دور؛ NDVI؛ LST. soil moisture, Remote Sensing, NDVI, LST 141 160 http://magazine.hormozgan.ac.ir/browse.php?a_code=A-10-986-1&slc_lang=fa&sid=1 mina eghtedarnezhad مینا اقتدارنژاد eghtedarnejad.m@stu.yazd.ac.ir 10031947532846008843 10031947532846008843 No Department of Rangeland and Watershed Management, Faculty of Natural Resources, Yazd University, Yazd. گروه مرتع و آبخیزداری، دانشکده منابع طبیعی و کویرشناسی، دانشگاه یزد، یزد Hossein Malekinezhad حسین ملکی‌نژاد hmalekinezhad@yazd.ac.ir 10031947532846008844 10031947532846008844 Yes Department of Rangeland and Watershed Management, Faculty of Natural Resources, Yazd University, Yazd. گروه مرتع و آبخیزداری، دانشکده منابع طبیعی و کویرشناسی، دانشگاه یزد، یزد Elham Rafiei sardooi الهام رفیعی ساردوئی ellrafiei@ujiroft.ac.ir 10031947532846008845 10031947532846008845 No Department of Ecological Engineering, Faculty of Natural Resources, University of Jiroft, Kerman. گروه مهندسی طبیعت، دانشکده مهندسی منابع طبیعی، دانشگاه جیرفت، جیرفت