| The actual production capacity of cultivated land is the basis for a series of farmland evaluations and is an important guarantee for food security issues.However,the current actual production capacity of arable land is still not objective and accurate.This study tries to improve the accuracy of the estimation of farmland production capacity.Zengcheng City in Guangdong Province was selected as the research area,and paddy fields and irrigated lands were selected as research objects,late season rice was set as the standard crop for the study.First,this paper develops a remote sensing mapping method for late season rice based on phenological characteristics and tests its accuracy.The spatial distribution of late season rice in the study area was combined with the three captured periods(around 1995,around 2005,and around 2015),and the feasibility and reliability of the research program to use late season rice as a standard crop were discussed.Secondly,above-ground biomass of late season rice was estimated using the EC-LUE model combined with the ESTARFM algorithm.The accuracy of the two models was verified and the above-ground biomass of late season rice in the study area in two periods(around 2005 and around 2015)was obtained.The actual production capacity differences and spatial distribution of the main cultivated land types in the study area were analyzed.Finally,combining the relevant materials and data of cultivated land surveys during the corresponding period,the various factors affecting the productive capacity of cultivated land were sorted out.Through statistical analysis methods such as correlation,statistics,and principal components,the main influencing factors of the main cultivated land productivity in the study area during the period of 2005 and 2015 were discussed,and the mechanism of the effect of single factors on farmland production capacity was discussed.The conclusions of this study are as follows:1.Through the remote sensing mapping method of late season rice based on key phonological phases,the spatial distribution of late season rice planting in the study period in the periods of 1995,2005 and 2015 was obtained.Inspections were conducted with field survey data and test samples made with higher spatial resolution images.The overall precision of late season rice mapping in the three periods was 91.0%,78.0%,and 90.0%,respectively.The user accuracy and cartographic accuracy of the cartographic results in 1995 and 2015 both reached 88.0%and above,while in 2005 the user accuracy and cartographic accuracy of cartographic results did not exceed 80.0%.Data quality and spatial resolution are considered to be the main factors affecting the accuracy of cartography.Since the 1990s,the cultivation area of late season rice in the study area has been greatly reduced.In 1995,the area of late season rice cultivation was263.43 km~2,and in 2005 it dropped to 133.69 km~2,and in 2015 it was reduced to less than 30 km~2.Late season rice cultivation has become fragmented,and more and more large and medium-sized rice fields have been transformed into small-scale rice fields.The average area of single late season rice fields has dropped from 8740.7 km~2 in 1995to 2292.3 km~2 in 2015.Through the analysis of late season rice mapping results and the existing farmland grading results and historical data of the study area,this study believes that adopting late season rice as a standard crop is fully representative of the long-term production capacity of the main cultivated land types in the study area.2.The results of cross-checking data fusion of key growth data of late season rice based on ESTARFM were Pearson coefficients of 0.651(2004292),0.646(2006265),0.428(2016263),and0.362(2013284)(all of which were statistically significant),and the fusion results were satisfactory.The late season rice area difference has no obvious effect on the ESTARFM algorithm.The results of estimating the biomass of late season rice on the EC-LUE model combined with the ESTARFM algorithm are:The average absolute error between predicted and measured biomass of late season rice in 2016 was240.35 g/m~2,standard error was 0.108,and Pearson coefficient was 0.630(statistically significant).By cross-checking the forecast results of two pairs of adjacent years,the correlation coefficient of the predicted above-ground biomass of late season rice in2016 and 2013 was 0.541(significance less than 0.001),and the data in 2006 and 2004was 0.606(significance is less than 0.001).The above-ground biomass of late season rice in the study area is predicted to be 1219 g/m~2 to 2840 g/m~2(2016),1435 g/m~2 to3079 g/m~2(2013),1042 g/m~2 to 2596 g/m~2(2006),1252 g/m~2 to 2997 g/m~2(2004).The estimates of above-ground biomass of late season rice in adjacent years are consistent in spatial distribution.3.Analyze the correlation between the estimated late-ground rice biomass and the farmland production capacity data collected during the corresponding period.During the period of 2005,there were certain correlations between the seven types of influencing factors of water source type,subsoil species,profile configuration,available phosphorus,available potassium,landform types,and irrigation patterns.The results of the correlation analysis for the 2015 period were five factors that influenced the type of water source,the type of soil it belongs to,the profile configuration,the type of landform,and the irrigation guarantee rate.4.The factors obtained from the correlation analysis were further analyzed with the above-ground biomass of late season rice for statistical analysis.In the 2005 period,the conditions of the water source were reservoirs,irrigation conditions were sprinkler irrigation,soil types were alluvial sand and mud fields,terrain was plains or hills,and soil profile conditions were A-B-C,with higher production capacity.Conversely,the arable land with low production capacity is irrigated with micro-irrigation or furrow irrigation,soil or vegetable land or other types,and the soil profile is A-C-C.In the2015 period,high-yielding arable land had the following conditions:the water source was a reservoir,the irrigation guarantee rate was Grade 1 or Grade 2,the landform condition was plain or hilly,the soil type was red muddy field,and the soil profile is loam-sand-loam or loam-clay-loam.The conditions for low-yielding arable land are as follows:the type of soil is white eel mud field,gully alluvial field or other types of field,and the soil profile is viscous.During the 2005 period,more suitable soil nutrients,irrigation methods,soil profiles,and soil types led to a higher performance of farmland productivity in the study area.Unsuitable soil nutrient conditions were the biggest impediment to rice paddy production capacity during this period.During the period of2015,more appropriate irrigation guarantee rates,soil types,and landform types were the guarantee for higher productivity of cultivated land.Unsuitable irrigation guarantee rates,soil types,landform types,and soil profile configurations also impede paddy field productivity. |
PDF Full Text Request